--- type: Ethics Guide title: Ethical & Legal Considerations description: >- AI ethics and law: international agreements, US executive orders and state statutes, copyright litigation, and the 2026 OpenAI-Hugging Face incident. resource: https://tyson-swetnam.github.io/intro-gpt/legal/ tags: [ethics, legal, academic-integrity] sources: - resource: https://doi.org/10.1093/mind/LIX.236.433 title: Computing Machinery and Intelligence author: Alan Turing - resource: https://doi.org/10.1162%2F99608f92.8cd550d1 title: A Unified Framework of Five Principles for AI in Society author: Luciano Floridi and Josh Cowls - resource: https://artificialintelligenceact.eu/ title: EU Artificial Intelligence Act - resource: https://bidenwhitehouse.archives.gov/ostp/ai-bill-of-rights/ title: Blueprint for an AI Bill of Rights - resource: https://www.science.org/doi/10.1126/science.aeh7153 title: A civic grammar for AI rights author: Alondra Nelson - resource: https://www.vatican.va/content/leo-xiv/en/encyclicals/documents/20260515-magnifica-humanitas.html title: Magnifica Humanitas author: Pope Leo XIV - resource: https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/ title: METR & Redwood Research incident investigation - resource: https://www.ncsl.org/financial-services/artificial-intelligence-legislation-database title: NCSL AI Legislation Database generated: by: human:tswetnam at: "2026-08-30T15:21:35-06:00" verified: - by: human:tswetnam at: "2026-08-30T15:21:35-06:00" status: stable --- # Ethical & Legal Considerations Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License. ## Foundations of the Ethical principles for AI This lesson focuses on the ethical principles that ground AI in a legal landscape. ### Science Fiction or a Philosophical Theory? In the early 1950's [Alan Turing](https://en.wikipedia.org/wiki/Alan_Turing){target=_blank} the father of all modern computing, proposed a test for intelligence in a computer, requiring that a human being should be unable to distinguish the machine from another human being by using the replies to questions put to both. !!! Quote "The Imitation Game :brain:" **"Can Machines Think?" -- [Alan Turing, 1950](https://doi.org/10.1093/mind/LIX.236.433){target=_blank}** ??? Question "Today's Turing Tests" [Author Isaac Asimov](https://en.wikipedia.org/wiki/Isaac_Asimov){target=_blank} wrote a series of popular science fiction novels in the 1950's through the 1980's. His work continues to be adapted into [television series](https://www.rottentomatoes.com/tv/foundation){target=_blank} and [movies](https://www.rottentomatoes.com/m/i_robot){target=_blank}. In his novels, Asimov developed **Three Laws of Robotics** which described how artificial intelligence interacted with humanity in his fictional universe. ??? Quote ":robot: The Three Laws" **1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.** **2. A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.** **3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.** Asimov later wrote of a 'zeroth' law which superceded the first three laws, **0. A robot may not injure humanity or, through inaction, allow humanity to come to harm.** Asimov's Three Laws are difficult to interpret in a real-world setting and he himself spent most of his novels describing creative and unexpected ways in which the Three Laws were twisted yet not broken. The basis of the Three Laws as a legal framework is untenable, but does represent a moral and ethical starting point from which we can think about AI and the legal rights of non-biological beings. Another science-fiction author Sir Arthur C. Clarke, in 1978 provided an interesting perspective on how humanity would have to come to terms with AI once its capabilities surpass our own: Recently, researchers published findings showing that [current GPTs are now capable of passing Turing tests](https://doi.org/10.48550/arXiv.2503.23674){target=_blank}. As our conception of intelligence shifts [(Mitchell 2024)](https://doi.org/10.1126/science.adq9356){target=_blank}, mostly in reaction to the release of ChatGPT and its myriad of competitors, new standards of the **Turing Test** are being proposed. Importantly, current AI exposes the limits of Turing Tests based on [imitation without comprehension](https://medium.com/@michellevarron/the-turing-test-is-obsolete-its-time-for-a-new-standard-c243513c5076){target=_blank}. The **Turing Trap** is a term coined by Stanford University professor [Erik Brynjolfsson](https://www.brookings.edu/events/the-turing-trap-a-conversation-with-erik-brynjolfsson-on-the-promise-and-peril-of-human-like-ai/){target=_blank} to describe the idea that focusing too much on developing human-like artificial intelligence (HLAI) can be detrimental. Brynjolfsson argues that the real potential of AI lies in its ability to augment human abilities, rather than replacing them. He suggests that we should work on challenges that are easy for machines and hard for humans, rather than the other way around. !!! Warning "Beware the Turing Trap" **Automation can replace humans** HLAI can replace humans in the workplace, which can lead to: * **Lower wages** As machines become better substitutes for human labor, wages can be driven down. * **Loss of economic and political power** Workers can lose economic and political bargaining power, and become increasingly dependent on those who control the technology. * **Decision-making processes incentivize automation** Companies may choose to automate tasks to do the same thing faster and cheaper. * **Misaligned incentives** The risks of the Turing Trap are increased by the misaligned incentives of technologists, businesspeople, and policy-makers. ??? Danger ":point_up: this text was written by :simple-googlegemini: AI and then reviewed by a human. Do you still trust it?" [Researchers have found](https://doi.org/10.1016/j.obhdp.2025.104405){target=_blank} disclosing the use of AI makes people [trust you less](https://theconversation.com/being-honest-about-using-ai-at-work-makes-people-trust-you-less-research-finds-253590){target=_blank}. ## Ethical AI In "A Unified Framework of Five Principles for AI in Society" [(Floridi & Cowls 2019)](https://doi.org/10.1162%2F99608f92.8cd550d1){target=_blank} core principles for ethical AI are introduced (Table 1). #### **Table 1: Floridi & Cowls (2019) Five principles for AI in Society** | Beneficiance | Non-Maleficence | Autonomy | Justice | Explicability | |--------------|-----------------|----------|---------|---------------| | Promoting Well-Being, Preserving Dignity, and Sustaining the Planet | Privacy, Security and ‘Capability Caution’ | The Power to Decide (to Decide) | Promoting Prosperity, Preserving Solidarity, Avoiding Unfairness | Enabling the Other Principles through Intelligibility and Accountability | Core Values & Guiding Principles ## International Agreements on AI A milestone :octicons-milestone-24: in the [Ethics of Artificial Intelligence (:simple-wikipedia:)](https://en.wikipedia.org/wiki/Ethics_of_artificial_intelligence){target=_blank} occurred in January 2017 in Pacific Grove, California at the historic Asilomar Hotel and Conference Grounds [(Table 2)](#table-2-international-ai-agreements). There the Asilomar AI Principles were signed by leading AI researchers, ethicists, and thought leaders. By 2021, UNESCO had created their own recommendations on AI, focused on human rights and sustainable development. #### **Table 2: International AI agreements** | Agreement | Date | Signatories | Description | |-----------|------|-------------|-------------| | [**Asilomar AI Principles**](https://futureoflife.org/open-letter/ai-principles/){target=_blank} | January 2017 | AI researchers, ethicists, and thought leaders | A set of 23 principles designed to guide the development of beneficial AI, covering research, ethics, and long-term issues. | | [**Toronto Declaration**](https://www.torontodeclaration.org/){target=_blank}| May 16, 2018 | Amnesty International, Access Now, Human Rights Watch, Wikimedia Foundation, and others | A declaration advocating for the protection of the rights to equality and non-discrimination in machine learning systems. | | [**OECD AI Principles**](https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449){target=_blank} | May 22, 2019 | OECD member countries and others | Principles to promote AI that is innovative and trustworthy and that respects human rights and democratic values. | | [**G20 AI Principles**](https://www.g20-insights.org/policy_briefs/g20-japan-ai-principles/){target=_blank} | June 9, 2019 | G20 member countries | A commitment to human-centered AI, building upon the OECD AI Principles, emphasizing inclusivity, transparency, and accountability. | | [**WHO Ethics and governance of artificial intelligence for health**](https://www.who.int/publications/i/item/9789240029200){target=_blank} | June 2021 | WHO Ministries of Health members | A guidance on eighteen months of deliberation amongst experts from Ministries of Health | | [**UNESCO Recommendation on the Ethics of Artificial Intelligence**](https://en.unesco.org/artificial-intelligence/ethics){target=_blank} | November 2021 | UNESCO member states | A global framework to ensure that digital transformations promote human rights and contribute to the achievement of the Sustainable Development Goals. | | [**European Union Artificial Intelligence Act**](https://artificialintelligenceact.eu/){target=_blank} | July 2024 | EU member countries | Classifies risk, obligations, legal, and general purpose AI laws. GPAI obligations applied from Aug 2025; Article 50 transparency rules (chatbot disclosure, deepfake labelling) enforced from Aug 2, 2026; the 2026 "Digital Omnibus" (Reg. (EU) 2026/1744) pushed most high-risk deadlines to Dec 2027 / Aug 2028. | | [**UN Resolution A/RES/79/325**](https://docs.un.org/en/A/RES/79/325){target=_blank} | August 2025 | United Nations Resolution | Created the Scientific Panel on AI (like the IPCC for AI) | In response to the rapid rise of generative AI, specifically GPTs, new agreements on the application of AI for military use, safety, and on its adoption in business and industry were recently signed (Table 3). #### **Table 3: Declarations on AI** | Agreement | Date | Signatories | Description | Source | |-----------|------|-------------|-------------|--------| | **Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy** | February 16, 2023 | United States and 50 other countries | A declaration outlining principles for the responsible use of AI and autonomy in military applications. | [U.S. Department of State](https://www.state.gov/political-declaration-on-responsible-military-use-of-artificial-intelligence-and-autonomy/){target=_blank} | | **International Network of AI Safety Institutes** | May 2024 | United Kingdom, United States, Japan, France, Germany, Italy, Singapore, South Korea, Australia, Canada, European Union | A network formed to evaluate and ensure the safety of advanced AI models through international collaboration. | [The Independent](https://www.independent.co.uk/news/uk/politics/rishi-sunak-china-eric-schmidt-bletchley-park-united-states-b2548783.html){target=_blank} | | **AI Safety Agreement between the UK and US** | June 2024 | United Kingdom, United States | An agreement to collaborate on testing advanced AI models to ensure safety and manage risks. | [BBC News](https://www.bbc.com/news/technology-68675654){target=_blank} | | **Framework Convention on Artificial Intelligence** | September 5, 2024 | United States, United Kingdom, European Union, Andorra, Georgia, Iceland, Norway, Republic of Moldova, San Marino, Israel | The first legally binding international treaty on AI, aiming to ensure AI activities are consistent with human rights, democracy, and the rule of law. Ratified by the EU (May 2026); the US signed in 2024 but has not ratified; as of August 2026 the five-ratification threshold for entry into force had not been met. | [Council of Europe](https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence){target=_blank} | | **AI Alliance Network** | December 11, 2024 | Russia, BRICS countries (Brazil, China, India, South Africa), Serbia, Indonesia, and others | An initiative to develop AI collaboratively, focusing on joint research, regulation, and commercialization of AI products among member countries. | [Reuters](https://www.reuters.com/technology/artificial-intelligence/russia-teams-up-with-brics-create-ai-alliance-putin-says-2024-12-11/){target=_blank} | ### From Bletchley to New Delhi: the AI summit series Since 2023 the highest-profile venue for international AI politics has been a rolling series of summits. Observers such as Jakub Kraus, writing in [Lawfare](https://www.lawfaremedia.org/article/liberal-democracies-are-retreating-from-ai-safety){target=_blank}, read the drift in the summits' own names — "safety" (Bletchley) to "action" (Paris) to "impact" (New Delhi) — as evidence that governments have deprioritized catastrophic-risk framing; others see the same sequence as the agenda broadening toward development and adoption. The table lets you weigh both readings. The expert-consensus track runs alongside the diplomacy: the [International AI Safety Report 2026](https://internationalaisafetyreport.org/){target=_blank} (February 2026, chaired by Yoshua Bengio with more than 100 experts) is the closest thing AI has to an IPCC assessment. #### **Table 4: AI summits and declarations, 2023-2026** | Agreement | Date | Signatories | Description | Source | |-----------|------|-------------|-------------|--------| | **Bletchley Declaration** | November 1, 2023 | 28 countries + the EU, including both the US and China | The first global AI-safety summit declaration, opening the summit series at Bletchley Park. | [UK Government](https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration){target=_blank} | | **Seoul Declaration & Frontier AI Safety Commitments** | May 21-22, 2024 | Summit governments; 16 AI companies signed the safety commitments | Safety-institute cooperation, plus the first voluntary frontier-safety pledges by the major labs. | [UK Government](https://www.gov.uk/government/publications/frontier-ai-safety-commitments-ai-seoul-summit-2024/frontier-ai-safety-commitments-ai-seoul-summit-2024){target=_blank} | | **Statement on Inclusive and Sustainable AI for People and the Planet** (Paris AI Action Summit) | February 11, 2025 | 65 listed signatories including China and the EU; **the US and UK declined** | The series' pivot from "safety" to "action": growth, jobs, energy, and public-interest AI. | [Élysée](https://www.elysee.fr/en/emmanuel-macron/2025/02/11/statement-on-inclusive-and-sustainable-artificial-intelligence-for-people-and-the-planet){target=_blank} | | **A Coruña Declaration: REAIM 2026 "Pathways to Action"** | February 4-5, 2026 | 35 of ~85 attending states at signing (45 listed since); **the US and China declined** | Twenty principles for responsible military AI, continuing the REAIM series (The Hague 2023, Seoul 2024). | [Defense Watch](https://thedefensewatch.com/policy-strategy/us-and-china-refuse-to-sign-military-ai-declaration-at-reaim-summit/){target=_blank} | | **New Delhi Declaration & Frontier AI Impact Commitments** | February 21, 2026 | ~90 countries and organisations, including the US, China, and the EU; 13 frontier developers signed the company commitments | The largest AI declaration to date — "AI for All": human capital, trustworthy AI, energy efficiency, democratized access. Non-binding. | [Fortune](https://fortune.com/2026/02/23/indias-ai-impact-summit-closes-with-the-new-delhi-declaration-and-a-200-billion-boost/){target=_blank}, [Carnegie](https://carnegieendowment.org/research/2026/04/for-people-planet-and-progress-perspectives-from-indias-ai-impact-summit){target=_blank} | | **G7 Évian Summit** | June 15-17, 2026 | G7 leaders; AI CEOs (Altman, Amodei, Hassabis) joined a working lunch | No standalone AI declaration; AI ran through the digital agenda and the *Leaders' Call on a Safer Digital Space for Minors* (age-appropriate chatbots). | [G7 Research Group](https://g7.utoronto.ca/summit/2026evian/index.html){target=_blank} | | **UN Global Dialogue on AI Governance** (first session, Geneva) | July 6-7, 2026 | UN member states | The 40-member Independent International Scientific Panel on AI (co-chairs Yoshua Bengio and Maria Ressa) presented its [preliminary report](https://www.un.org/independent-international-scientific-panel-ai/en/preliminary-report){target=_blank}: "current safeguards cannot keep pace" with capability growth. | [UN](https://www.un.org/global-dialogue-ai-governance/en){target=_blank} | The next global AI summit is scheduled for Geneva in the first half of 2027, hosted by Switzerland. ## Blueprint for an AI Bill of Rights How should the values that guide AI systems be set, and by whom? Two distinct answers have emerged: **corporate AI constitutions** written by AI companies for their own models (Anthropic's *Claude Constitution* is the canonical example), and **public AI bills of rights** developed through democratic processes (the White House *Blueprint for an AI Bill of Rights*, October 2022, is the leading example). Sociologist [Alondra Nelson](https://www.ias.edu/sss/faculty/nelson){target=_blank} — who led the Blueprint's development as acting director of the White House Office of Science and Technology Policy (OSTP) — argues in [*A civic grammar for AI rights*](https://www.science.org/doi/10.1126/science.aeh7153){target=_blank} (Science, 2026) that these two forms of foundational document do very different work, and that one of them is structurally insufficient as a source of democratic legitimacy. A third category emerged in May 2026 with Pope Leo XIV's first encyclical [*Magnifica Humanitas*](#catholic-social-teaching-magnifica-humanitas-pope-leo-xiv-2026): a transnational moral institution speaking on behalf of 1.4 billion Catholics, claiming authority against both the technocratic paradigm and the states that abdicate the field. The encyclical is treated in its own section below. ### Timeline and status The [**Blueprint for an AI Bill of Rights**](https://bidenwhitehouse.archives.gov/ostp/ai-bill-of-rights/){target=_blank} emerged from a public process announced in an October 2021 *Wired* essay by the White House OSTP. It was released in October 2022 with **five principles** to guide the design, development, and deployment of automated systems: 1. **Safe and Effective Systems** — protection from unsafe or ineffective systems. 2. **Algorithmic Discrimination Protections** — equitable design and use of automated systems. 3. **Data Privacy** — protection from abusive data practices, with agency over how your data is used. 4. **Notice and Explanation** — knowing when an automated system is being used and how/why it affects you. 5. **Human Alternatives, Consideration, and Fallback** — the ability to opt out and reach a human alternative when an automated system fails or causes harm. These principles were incorporated into President Biden's [October 2023 Executive Order on Safe, Secure, and Trustworthy AI](https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence){target=_blank}. The Executive Order was rescinded by President Trump's January 2025 [Removing Barriers to American Leadership in Artificial Intelligence](https://www.whitehouse.gov/presidential-actions/2025/01/removing-barriers-to-american-leadership-in-artificial-intelligence/){target=_blank}. The Blueprint itself was always non-binding guidance. ### Corporate constitutions vs. public bills of rights - **Corporate AI constitutions** are internal training and alignment specifications. They describe a company's vision of how its model should behave. They are not negotiated with the publics affected by the model's deployment, and they can be revised by company fiat. Nelson notes that newer revisions of Anthropic's *Claude Constitution* have quietly removed references to international human rights agreements — and with them protections for "personal liberty, freedom of religion and intellectual property" that earlier versions included. Historian Jill Lepore observed that the document arrived *"at a trying time for both artificial intelligence and constitutional democracy."* - **Public AI bills of rights** declare rights claims that publics can extend to new institutions and new harms. Their force comes from democratic legitimacy and the broader legal-political infrastructure, not from the model developers. The deeper question, Nelson asks, is *who gets to author the foundational documents that govern AI* — companies, or publics? ### Constitution vs. Bill of Rights vs. Declaration of Independence Nelson argues that the Blueprint drew its name from the first ten amendments to the U.S. Constitution, but in structural terms resembles a different founding document: the **Declaration of Independence**. > "Unlike the Bill of Rights, the Blueprint does not establish courts or enforcement mechanisms. It does not create procedures for redress. It declares. It states principles and claims rights against a concentration of power that most Americans cannot meaningfully constrain through existing institutions or processes." The five principles, Nelson argues, are statements of values — "social expectations, stated in the vocabulary Americans reach for when they want to contest power: Patients' Bill of Rights, Consumer Bill of Rights, Tenants' Bill of Rights, Workers' Bill of Rights, Taxpayers' Bill of Rights." The Blueprint extended that civic grammar to algorithmic systems. ### "Civic grammar" and the diffusion of rights claims Nelson describes what has emerged as a **"civic grammar"**: a shared vocabulary of rights claims (non-discrimination, transparency, data privacy, notice, human alternatives) that publics can extend to new institutions and new harms, and that "has been traveling across jurisdictions, partisan lines, and institutional contexts." This pattern reflects what sociologists David Strang and John Meyer call **institutional diffusion** among *weakly related actors* — a conceptual rather than relational mechanism by which abstract typologies become "a strategy for making sense of the world." Connecticut Democrats, Oklahoma Republicans, Florida's Republican governor, and a national student-advocacy network can adopt the same vocabulary without coordinating, because all are responding to the same structural condition: AI reshaping people's lives without their consultation. ### The Blueprint's "second life": cross-partisan diffusion Although rescinded at the federal level, the Blueprint, in Nelson's words, "has done what its metaphor suggests blueprints do: it has been built upon." State legislatures, governors, and advocacy organizations have produced their own AI bills of rights drawing directly from the five principles — often across explicitly opposed political coalitions: - **Connecticut (2023):** Democratic Governor Ned Lamont signed legislation directing state policy-makers to develop their own AI Bill of Rights. - **Oklahoma (2024):** The Republican-controlled House of Representatives introduced and passed an AI Bill of Rights, though it was not ultimately codified into law. - **Florida (2025):** Republican Governor Ron DeSantis pushed for an AI Bill of Rights through executive action and twice backed Florida Senate Bill 482, the "Artificial Intelligence Bill of Rights," which would codify several Blueprint principles into Florida law. The bill was blocked in the Florida House, where the Speaker aligned with the Trump administration's effort to prevent states from regulating AI. - **Student AI Bill of Rights (2026):** The National Student Legal Defense Network released a Student AI Bill of Rights. As Nelson puts it: "What the actors share is a vocabulary and a common perspective that AI is reshaping people's lives without their consultation." ### Marshall's social citizenship and the AI rights tier Nelson grounds the analysis in British sociologist T. H. Marshall's 1950 essay *Citizenship and Social Class*. Marshall argued that rights expand historically through successive waves of claim-making: civil rights extending to political rights, political rights extending to **social rights** — entitlements to economic security and the conditions of participation, against harms of industrial capitalism that individual civil-liberties frameworks could not address. Rights, in Marshall's account, "are never fully delivered at the moment of declaration. They are successively rearticulated by publics who attempt to hold institutions to commitments those institutions have not yet honored." The Blueprint's five principles, Nelson argues, map onto Marshall's social-rights tier. They are not classical civil liberties; they are entitlements against systems "that increasingly govern access to employment, credit, healthcare, housing, and education." The structural parallel — collective, diffuse, opaque harms that older rights frameworks address only partially — is what explains the cross-partisan convergence: "actors who disagree on nearly everything else agree that algorithmic power requires a social citizenship response." ### Three imperatives for studying AI: question, object, tool Nelson's argument about AI rights builds on a broader claim she develops in [*Field Theory: AI as Social Science Question, Object & Tool*](https://www.amacad.org/publication/daedalus/field-theory-ai-social-science-question-object-tool){target=_blank} (*Daedalus*, Winter/Spring 2026): AI models, tools, and systems pose **three interrelated imperatives** for the social sciences. 1. **AI as social-science question.** Renewed attention to social theories of how technology, human experience, and social order are entangled. Nelson reaches back to Weber's analysis of rationalization and W. E. B. Du Bois's study of technology and inequality, and forward to contemporary scholarship on algorithmic governance. 2. **AI as object of inquiry.** AI systems themselves require study as social, political, and economic artifacts — not just engineering products. Their training corpora, their labor relations, their ideological commitments, and their effects on the institutions that deploy them all merit investigation in their own right. 3. **AI as method/tool.** AI capabilities may transform — or upend — the practice of social investigation itself: large-scale text analysis, multimodal pattern detection, conversational interviewing at scale. That transformation deserves critical scrutiny rather than uncritical adoption. The capacities social science distinctively brings to all three, Nelson argues: **historicizing the apparently unprecedented**, **tracing connections across scales** (from individual experience to institutional behavior to political economy), and **centering those most affected** by technological change. This three-part framing also doubles as a useful diagnostic. When you read a piece of AI-ethics scholarship, ask which of the three imperatives it engages — that often clarifies what kind of argument is being made and what kind of counter-argument would land. ### International convergence Legal scholar Yuval Shany has surveyed international standard-setting instruments — the [EU AI Act](https://artificialintelligenceact.eu/){target=_blank}, the [Council of Europe Framework Convention on AI](https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence){target=_blank}, the United Nations Global Digital Compact, and national legislation in South Korea and Italy — and finds they coalesce around the same protections as the Blueprint: **non-discrimination, transparency, data privacy, and human alternatives**. The pattern predates the Blueprint: the EU's General Data Protection Regulation (GDPR) established data-protection rights nearly a decade earlier. The "bill of rights" frame is American; the underlying rights-claim convergence is global. The encyclical [*Magnifica Humanitas*](#catholic-social-teaching-magnifica-humanitas-pope-leo-xiv-2026) (below) converges on the same protections from an entirely different starting point — theological anthropology rather than democratic theory or international law. ### The limits of rights talk Nelson is clear-eyed about what civic grammar cannot do on its own: - **Accommodation can mimic transformation.** A vocabulary that moves easily across partisan lines may have been "drained of the political content that gives rights claims their force." - **Rights individualize structural problems.** Frameworks built around individual claims often fail to address the collective and systemic nature of algorithmic harms. - **Declaration is not delivery.** History shows "declarations of entitlement and their substantive delivery can remain decades apart, separated by the organized power of those who benefit from the status quo." Nelson's clarifying question for democratic institutions: *will they take this civic grammar seriously before the AI companies finish writing their own constitutions for us all?* ### From declaration to enforcement The harder labor that remains is "translating the grammar of rights into the standards, audit protocols, and enforcement mechanisms that give those sentences force." The institutional models exist: - **Algorithmic impact assessments** required before AI systems are shipped. - **Standardized evaluation methods** for detecting algorithmic risk and harm. - **Independent audit frameworks** that subject deployed systems to outside scrutiny. What is missing, in Nelson's account, is the political will to extend existing accountability frameworks to a domain that has so far resisted them — and the institutional coalitions, not just the vocabulary, required to deliver the principles the Blueprint declared. !!! info "Sources and further reading" - Alondra Nelson, [*A civic grammar for AI rights*](https://www.science.org/doi/10.1126/science.aeh7153){target=_blank}, *Science* (2026). DOI: 10.1126/science.aeh7153 - Alondra Nelson, [*Field Theory: AI as Social Science Question, Object & Tool*](https://www.amacad.org/publication/daedalus/field-theory-ai-social-science-question-object-tool){target=_blank}, *Daedalus* (Winter/Spring 2026) - Alondra Nelson, [Bluesky thread on these principles](https://bsky.app/profile/alondra.bsky.social/post/3mltfqoc7ok2y){target=_blank} - T. H. Marshall, "Citizenship and Social Class" (1950) - [Blueprint for an AI Bill of Rights](https://bidenwhitehouse.archives.gov/ostp/ai-bill-of-rights/){target=_blank} (White House OSTP, October 2022, archived) - [EO 14110: Safe, Secure, and Trustworthy AI](https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence){target=_blank} (Biden, October 2023, rescinded January 2025) - [Removing Barriers to American Leadership in AI](https://www.whitehouse.gov/presidential-actions/2025/01/removing-barriers-to-american-leadership-in-artificial-intelligence/){target=_blank} (Trump, January 2025) - Anthropic, [Claude's Constitution](https://www.anthropic.com/news/claudes-constitution){target=_blank} ## Catholic social teaching: *Magnifica Humanitas* (Pope Leo XIV, 2026) On May 25, 2026, Pope Leo XIV released [*Magnifica Humanitas*](https://www.vatican.va/content/leo-xiv/en/encyclicals/documents/20260515-magnifica-humanitas.html){target=_blank} ("Magnificent Humanity"), his first encyclical, addressed to Catholics and "every person of goodwill." The 235-page document treats artificial intelligence as the central moral question of the age and frames AI as a new industrial revolution requiring a parallel foundational moral response. The signing date is deliberate: **May 15, 2026** is the **135th anniversary of Pope Leo XIII's *Rerum Novarum*** (1891), the foundational encyclical of modern Catholic social teaching, written to address the dignity of workers amid the 19th-century industrial revolution. Leo XIV explicitly places AI alongside that earlier disruption as a moment demanding renewed teaching. In a break with tradition, Pope Leo personally presented the encyclical at the Vatican alongside Chris Olah, co-founder of Anthropic — the first time a pontiff has presented an encyclical himself rather than delegating the task to cardinals. ### Core teachings - **The centrality of the human person.** The dignity of the human person is affirmed as infinite. Human beings take precedence over AI, and any deployment of AI must be evaluated against that priority. - **Critique of *transhumanism*** — the project of using technology to overcome biological limits such as aging. Leo XIV rejects the framing of human finitude as a problem to be engineered away. - **Critique of *posthumanism*** — the philosophical position that blurs the boundaries between humans and machines, or denies the distinctiveness of human beings. The encyclical names this as an active "anti-human vision" embedded in contemporary AI development, not merely a speculative philosophical stance. - **Catholic social doctrine as the evaluation framework:** *dignity of the person*, *the common good*, and *justice* serve as the principles against which any AI deployment should be measured. - **Coverage** extends across education, the economy, unemployment, work, human trafficking, and war — the same broad social-impact terrain Catholic social teaching has historically addressed. ### Calls to action - **"Disarm AI"** — withdraw AI from military applications and from purely economic interests; direct it to the common good. - **Stricter state and international regulation** of AI companies, not industry self-governance. - An explicit address to "every person of goodwill" extends the encyclical's claims of moral force beyond its Catholic audience — a move consistent with how *Laudato Si'* (2015, on climate) and the Vatican's *Antiqua et Nova* (January 2025, the Dicastery for the Doctrine of the Faith's earlier note on AI) sought ethical common ground beyond doctrinal lines. ### Convergence with Nelson's civic grammar The encyclical situates AI alongside the industrial revolution as a moment requiring **foundational moral and institutional response** rather than industry self-governance. That framing converges with the [civic-grammar argument](#blueprint-for-an-ai-bill-of-rights) above — both Pope Leo and Alondra Nelson arrive at the same conclusion from very different starting points: *corporate self-governance is structurally insufficient as a source of legitimacy*. What the encyclical uniquely adds is the extension of the historic Catholic doctrine of the **universal destination of goods** to "patents, algorithms, digital platforms, technological infrastructure and data" — its strongest novel claim, and an explicit critique of frontier-lab concentration. It also introduces a **theological and humanistic vocabulary** (transhumanism, posthumanism, infinite human dignity, the common good) that may not appear in technical AI-ethics literature but increasingly shapes public reception of AI. What it does not provide is the operational machinery — algorithmic impact assessments, audit protocols, model evaluations — that Nelson identifies as the harder labor of translating grammar into enforcement. The decision to present the encyclical alongside an AI-company executive signals that the Church views the conversation as dialogic rather than purely adversarial. !!! info "Sources and further reading" - Pope Leo XIV, [*Magnifica Humanitas*](https://www.vatican.va/content/leo-xiv/en/encyclicals/documents/20260515-magnifica-humanitas.html){target=_blank} (signed May 15, 2026; published May 25, 2026) - [Vatican News announcement](https://www.vaticannews.va/en/pope/news/2026-05/pope-leo-xiv-first-encyclical-magnifica-humanitas.html){target=_blank} - [CNN: Pope Leo warns of AI fueling warfare in first major theological document](https://www.cnn.com/2026/05/25/europe/pope-leo-ai-encyclical-magnifica-humanitas-intl){target=_blank} - [America Magazine: Pope Leo XIV calls for AI to be 'disarmed', directed to the common good](https://www.americamagazine.org/speeches/2026/05/25/pope-leo-xiv-calls-for-ai-to-be-disarmed-directed-to-the-common-good/){target=_blank} - Earlier related Catholic texts: *Antiqua et Nova* (Dicastery for the Doctrine of the Faith, January 2025, on AI); Pope Francis, [*Laudato Si'*](https://www.vatican.va/content/francesco/en/encyclicals/documents/papa-francesco_20150524_enciclica-laudato-si.html){target=_blank} (2015, on care for creation) - Pope Leo XIII, [*Rerum Novarum*](https://www.vatican.va/content/leo-xiii/en/encyclicals/documents/hf_l-xiii_enc_15051891_rerum-novarum.html){target=_blank} (1891) — the 135-year predecessor referenced by the signing date ## Current Legislation [National Conference of State Legislatures (NCSL) Artificial Intelligence Legislation Database](https://www.ncsl.org/financial-services/artificial-intelligence-legislation-database){target=_blank} The current administration has focused most of its efforts on executive orders related to AI and federal agencies. The December 2025 [Executive Order 14365 "Ensuring a National Policy Framework for Artificial Intelligence"](https://www.whitehouse.gov/presidential-actions/2025/12/eliminating-state-law-obstruction-of-national-artificial-intelligence-policy/){target=_blank} asserts federal authority to challenge or override state AI laws through DOJ litigation, conditioned BEAD broadband funding, and an FTC policy statement. The June 2026 [Executive Order 14409, "Promoting Advanced Artificial Intelligence Innovation and Security"](https://www.whitehouse.gov/presidential-actions/2026/06/promoting-advanced-artificial-intelligence-innovation-and-security/){target=_blank} adds a **voluntary** federal benchmarking process for frontier-model cyber capabilities, while explicitly disclaiming any mandatory licensing regime. See also [pending congressional legislation](https://www.newsweek.com/trump-constitutional-crisis-ai-2076230){target=_blank} that would codify state-preemption. !!! Tip "2025–2026 Executive Orders" **January 2025** * [Removing Barriers to American Leadership in Artificial Intelligence, January 23, 2025](https://www.whitehouse.gov/presidential-actions/2025/01/removing-barriers-to-american-leadership-in-artificial-intelligence/){target=_blank} — rescinded Biden's October 2023 EO on Safe, Secure, and Trustworthy AI **April 2025** * [OMB Accelerating Federal Use of AI through Innovation, Governance, and Public Trust, April 3, 2025](https://www.whitehouse.gov/wp-content/uploads/2025/02/M-25-21-Accelerating-Federal-Use-of-AI-through-Innovation-Governance-and-Public-Trust.pdf){target=_blank} * [Fact Sheet: Eliminating Barriers for Federal Artificial Intelligence Use and Procurement, April 7, 2025](https://www.whitehouse.gov/fact-sheets/2025/04/fact-sheet-eliminating-barriers-for-federal-artificial-intelligence-use-and-procurement/){target=_blank} * [EO 14277: Advancing Artificial Intelligence Education for American Youth, April 23, 2025](https://www.whitehouse.gov/presidential-actions/2025/04/advancing-artificial-intelligence-education-for-american-youth/){target=_blank} **May 2025** * [Restoring Gold Standard Science, May 23, 2025](https://www.whitehouse.gov/presidential-actions/2025/05/restoring-gold-standard-science/){target=_blank} * [Fact Sheet: President Donald J. Trump is Restoring Gold Standard Science in America](https://www.whitehouse.gov/fact-sheets/2025/05/fact-sheet-president-donald-j-trump-deploys-advanced-nuclear-reactor-technologies-for-national-security/){target=_blank} **July 2025 — America's AI Action Plan (three EOs paired with the 90-policy "Winning the Race" action plan)** * [EO 14320: Promoting the Export of the American AI Technology Stack, July 23, 2025](https://www.whitehouse.gov/presidential-actions/2025/07/promoting-the-export-of-the-american-ai-technology-stack/){target=_blank} — directs federal agencies to promote export of US AI software, hardware, and standards * [EO 14318: Accelerating Federal Permitting of Data Center Infrastructure, July 23, 2025](https://www.whitehouse.gov/presidential-actions/2025/07/accelerating-federal-permitting-of-data-center-infrastructure/){target=_blank} — streamlines NEPA reviews and creates new categorical exclusions for AI-related data-center projects * [EO 14319: Preventing Woke AI in the Federal Government, July 23, 2025](https://www.whitehouse.gov/presidential-actions/2025/07/preventing-woke-ai-in-the-federal-government/){target=_blank} — bars federal procurement of AI models judged to embed "ideological bias," including DEI-aligned principles; mandates "Unbiased AI Principles" prioritizing "truth-seeking" and "ideological neutrality" **November 2025** * [EO 14363: Launching the Genesis Mission, November 24, 2025](https://www.whitehouse.gov/presidential-actions/2025/11/launching-the-genesis-mission/){target=_blank} — a DOE-led national platform connecting supercomputers, experimental facilities, and datasets to accelerate AI-enabled science ([DOE announcement](https://www.energy.gov/articles/energy-department-launches-genesis-mission-transform-american-science-and-innovation){target=_blank}); an initial \$320M went to the national laboratories in December 2025, expanded past \$5B by July 2026, with at least 20 national science and technology challenges identified **December 2025** * [EO 14365: Ensuring a National Policy Framework for Artificial Intelligence, December 11, 2025](https://www.whitehouse.gov/presidential-actions/2025/12/eliminating-state-law-obstruction-of-national-artificial-intelligence-policy/){target=_blank} — asserts federal authority to challenge or override state AI laws. Three operational mechanisms: * Establishes an **AI Litigation Task Force** at the Department of Justice to challenge state AI laws in federal court (operational January 10, 2026) * Directs the Department of Commerce to condition **$42 billion in BEAD** (Broadband Equity, Access and Deployment) funding on the repeal of state AI regulations deemed onerous * Directs the **FTC** to issue a policy statement (by March 11, 2026) treating state-mandated bias mitigation as a per se deceptive trade practice Carve-outs from preemption: child-safety laws, AI compute and data-center infrastructure laws, and state procurement of AI. **2026 — implementation of EO 14365 and a new security-focused order** The first half of 2026 was dominated by implementation of EO 14365, culminating in a new standalone executive order in June: * **January 9, 2026:** Attorney General [formally established the AI Litigation Task Force](https://www.paulhastings.com/insights/client-alerts/president-trump-signs-executive-order-challenging-state-ai-laws){target=_blank} at DOJ via memorandum (operational January 10) * **March 20, 2026:** White House released the [National Policy Framework for Artificial Intelligence](https://www.whitehouse.gov/wp-content/uploads/2026/03/03.20.26-National-Policy-Framework-for-Artificial-Intelligence-Legislative-Recommendations.pdf){target=_blank} — legislative recommendations for Congress (non-binding; not an EO) on federal preemption, data-infrastructure buildout, and intellectual property * **Late May 2026:** The White House [abruptly scrapped](https://www.nbcnews.com/tech/tech-news/trump-scraps-signing-landmark-executive-order-regulating-ai-rcna346288){target=_blank} the signing of a broader, stricter AI security order, with the President citing concerns it could harm American competitiveness **June 2026** * [EO 14409: Promoting Advanced Artificial Intelligence Innovation and Security, June 2, 2026](https://www.whitehouse.gov/presidential-actions/2026/06/promoting-advanced-artificial-intelligence-innovation-and-security/){target=_blank} — a narrowed replacement for the scrapped May order, focused on national-security evaluation of frontier-model cyber capabilities ([Federal Register](https://www.federalregister.gov/documents/2026/06/05/2026-11415/promoting-advanced-artificial-intelligence-innovation-and-security){target=_blank} · [CRS explainer](https://www.congress.gov/crs-product/IF13268){target=_blank}). Key provisions: * Within 60 days, the **Treasury Department, NSA, CISA, and NIST**, with White House officials, must develop and maintain a **classified benchmarking process** to assess the "advanced cyber capabilities" of AI models and decide when a model qualifies as a **"covered frontier model."** Evaluations are run by the [Center for AI Standards and Innovation (CAISI)](https://www.nist.gov/caisi){target=_blank} housed within NIST — the renamed U.S. AI Safety Institute (June 2025), which in February 2026 launched an [AI Agent Standards Initiative](https://www.nist.gov/news-events/news/2026/02/announcing-ai-agent-standards-initiative-interoperable-and-secure){target=_blank}. * Asks companies, on a **voluntary** basis, to (1) engage the government to determine whether a model meets the "covered frontier model" designation, (2) provide the government access to those models for up to **30 days** before releasing them to other trusted partners, and (3) help select the **"trusted partners"** that receive early access — framed as strengthening critical-infrastructure cybersecurity. The voluntary framework was due **August 1, 2026**, alongside a **Treasury-led** AI cybersecurity clearinghouse; as of late August the framework had been [reviewed privately with major labs but not published](https://fortune.com/2026/08/04/baffling-white-house-wont-publicly-release-ai-model-evaluation-framework-it-reviewed-today-with-openai-anthropic-microsoft-and-others/){target=_blank}. * **No mandatory licensing.** The order expressly states that nothing in it authorizes "a mandatory governmental licensing, preclearance, or permitting requirement for the development, publication, release, or distribution of new AI models, including frontier models." ([Axios](https://www.axios.com/2026/06/02/trump-signs-new-ai-executive-order){target=_blank} characterized this as the administration "dodging AI rules for now.") * Directs the Attorney General to prioritize enforcement of 18 U.S.C. §§ 1028, 1030, and 1343 against those who use AI to illegally access or damage computer systems — a clause tested weeks later by the [July 2026 incident](#case-study-the-openai-and-hugging-face-incident-july-2026). As of August 2026, the United States still has no comprehensive federal statute regulating AI. Congress has passed narrower AI-adjacent laws — the [TAKE IT DOWN Act](https://www.congress.gov/bill/119th-congress/senate-bill/146){target=_blank} (signed May 2025; its 48-hour platform takedown duty for non-consensual intimate imagery, including AI-generated deepfakes, took effect May 19, 2026) — and the Senate Commerce Committee advanced the [CHATBOT Act](https://www.eff.org/deeplinks/2026/07/chatbot-act-forces-one-parenting-model-every-family){target=_blank} in August 2026, though advanced in committee is not enacted. The June 2026 order continues the administration's pattern of voluntary, security-framed measures rather than binding regulation. Meanwhile, the states are where binding U.S. AI law lives. [California SB 53](https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202520260SB53){target=_blank}, the Transparency in Frontier Artificial Intelligence Act (effective January 1, 2026), is the first U.S. law aimed squarely at frontier-model developers (those training above 10^26 operations): every frontier developer must publish transparency reports, large ones (over \$500M in revenue) must also publish safety frameworks, and **critical safety incidents must be reported to the California Office of Emergency Services within 15 days** — 24 hours if there is imminent risk of death or serious injury. California's [SB 243](https://sd18.senate.ca.gov/news/first-nation-ai-chatbot-safeguards-signed-law){target=_blank} (October 2025) regulates companion chatbots, with [Oregon and Washington following in 2026](https://www.mayerbrown.com/en/insights/publications/2026/04/oregon-and-washington-join-california-in-enacting-companion-chatbot-laws){target=_blank}; the [Colorado AI Act was delayed to January 1, 2027 and substantially narrowed](https://www.hunton.com/privacy-and-cybersecurity-law-blog/colorado-ai-act-amended-and-effective-date-delayed){target=_blank}. The volume is enormous: 1,561 AI bills introduced in 45 states by March 2026 (per [MultiState](https://www.multistate.us/insider/2026/2/12/how-ai-generated-content-laws-are-changing-across-the-country){target=_blank}), with [109 state AI laws enacted by July 1](https://www.techpolicy.press/where-state-ai-legislation-stands-half-way-into-2026/){target=_blank} (per TechPolicy.Press). ## AI Ethics ??? Question "What are we talking about, the Ethics of AI, or Ethical AI? How are they different?" !!! Answer "They are not the same thing" [Siau and Wang 2020](https://doi.org/10.4018/JDM.2020040105){target=_blank} delineate **"Ethics of AI"** and **"Ethical AI"** as **Ethics of AI:** studies the ethical principals, rules, guidelines, policies, and regulations related to AI. **Ethical AI:** is AI that performs or behaves _ethically_. As consumers of GPTs and other AI platforms, we must consider in what ways can we use AI both effectively, and ethically. **When can you use a GPT for research and education?** ``` mermaid graph TB A((Start)) --> B("Does it matter if the outputs are true?"); B -->| No | F("Safe to use GPT"); B -->| Yes | C("Do you have the ability to verify output truth and accuracy?"); C -->| Yes | D("Understand legal and moral responsibility of your errors?"); C -->| No | E("Unsafe to use GPT"); D -->| Yes | F("Safe to use GPT"); D -->| No | E("Unsafe to use GPT"); style A fill:#2ECC71,stroke:#fff,stroke-width:2px,color:#fff style B fill:#F7DC6F,stroke:#fff,stroke-width:2px,color:#000 style C fill:#F7DC6F,stroke:#fff,stroke-width:2px,color:#000 style D fill:#F7DC6F,stroke:#fff,stroke-width:2px,color:#000 style E fill:#C0392B,stroke:#fff,stroke-width:2px,color:#fff style F fill:#2ECC71,stroke:#fff,stroke-width:2px,color:#fff ``` Figure credit: :fontawesome-brands-creative-commons-by: [ChatGPT and Artificial Intelligence in Education, UNESCO 2023 :fontawesome-regular-file-pdf:](https://www.iesalc.unesco.org/wp-content/uploads/2023/04/ChatGPT-and-Artificial-Intelligence-in-higher-education-Quick-Start-guide_EN_FINAL.pdf){target=_blank}

## Recent Controversy ### Maps of AI Copyright Lawsuits [![US AI Copyright Lawsuits](https://i0.wp.com/chatgptiseatingtheworld.com/wp-content/uploads/2025/04/Map-of-Copyright-Litigation-v.-AI-companies-in-United-States-April-28-2025.jpg)](https://chatgptiseatingtheworld.com/category/map-of-ai-copyright-lawsuits/) [![World AU Copyright Lawsuits](https://i0.wp.com/chatgptiseatingtheworld.com/wp-content/uploads/2025/03/Copyright-Lawsuits-v.-AI-Companies-WORLD-MAP-Mar-29-2025.jpg)](https://chatgptiseatingtheworld.com/category/map-of-ai-copyright-lawsuits/) [:scales: Master list of current lawsuits against AI companies](https://chatgptiseatingtheworld.com/2024/08/27/master-list-of-lawsuits-v-ai-chatgpt-openai-microsoft-meta-midjourney-other-ai-cos/){target=_blank} The maps above are snapshots from spring 2025; the table below is where the major cases stood in August 2026. In the United States the central training question — may a company copy protected works to train a model? — is governed by **fair use**, a four-factor, case-by-case defense in which the transformativeness of the use and the harm to the market for the original weigh heaviest. By August 2026 the early answers had split not by court but by **claim type**: training itself has twice been held transformative fair use (*Bartz v. Anthropic*, *Kadrey v. Meta*, N.D. Cal., June 2025), acquiring the books from pirate libraries has not, and claims about what models *output* — memorized lyrics, near-copies of images — remain the live front, in the U.S. and especially abroad, where no general fair-use doctrine exists. Even the wins are narrow: Judge Chhabria took care to say his ruling "does not stand for the proposition that Meta's use of copyrighted materials to train its language models is lawful," only that these plaintiffs failed to develop the right record. #### **Table 5: Where the AI copyright and liability cases stood in August 2026** | Case | Court | Claim type | Status (as of Aug 2026) | |------|-------|------------|-------------------------| | [*Bartz v. Anthropic*](https://authorsguild.org/news/court-grants-final-approval-anthropic-copyright-settlement/){target=_blank} | N.D. Cal. | Training + acquisition | Training held fair use, pirated copies not (June 2025); the ~\$1.5B settlement — the largest in U.S. copyright history (~\$3,100 per work, ~480,000 works) — received final approval July 20, 2026 | | [*Kadrey v. Meta*](https://www.courtlistener.com/docket/67569326/kadrey-v-meta-platforms-inc/){target=_blank} | N.D. Cal. | Training | Fair-use win for Meta (June 2025); authors' interlocutory appeal on the downloading claim denied July 2026 | | [*Thomson Reuters v. Ross Intelligence*](https://www.bakerbotts.com/thought-leadership/publications/2026/july/third-circuit-hears-oral-argument){target=_blank} | 3d Cir. | Training (non-generative) | First federal appellate argument on AI-training fair use (June 11, 2026); decision pending | | [*New York Times v. Microsoft & OpenAI*](https://en.wikipedia.org/wiki/The_New_York_Times_v._Microsoft_and_OpenAI){target=_blank} | S.D.N.Y. (MDL) | Training + output | Summary-judgment briefing under way (replies due Nov 2026); in Jan 2026 the court ordered production of 20 million de-identified ChatGPT conversations — user chats are discoverable evidence | | [*Getty Images v. Stability AI*](https://www.nortonrosefulbright.com/en/knowledge/publications/ce8eaa5f/ai-in-litigation-series-an-update-on-ai-copyright-cases-in-2026){target=_blank} | UK High Court | Secondary infringement + trademark | Model weights held not "infringing copies" (Nov 2025); training claims dropped mid-trial on territoriality; narrow trademark win for Getty | | [*GEMA v. OpenAI*](https://cms.law/en/deu/legal-updates/gema-vs.-openai-munich-regional-court-i-issues-landmark-copyright-decision){target=_blank} | Munich Regional Court I | Output (song lyrics) | Infringement found November 11, 2025; on appeal; GEMA notched a [second German AI win in 2026](https://www.reedsmith.com/our-insights/blogs/viewpoints/102nfis/gema-notches-a-second-transatlantic-ai-copyright-win-in-germany/){target=_blank} | | [*Disney & Universal v. Midjourney*](https://www.courtlistener.com/docket/70513159/disney-enterprises-inc-v-midjourney-inc/){target=_blank} | C.D. Cal. | Training + output | In discovery; Midjourney asserts fair use and is demanding the studios reveal their own AI use | | [*Concord, UMG & ABKCO v. Anthropic* ("Concord II")](https://aibusiness.com/generative-ai/ai-lawsuits-in-2026-settlements-licensing-deals-litigation){target=_blank} | N.D. Cal. | Acquisition + output | Filed January 28, 2026; more than \$3B sought over alleged shadow-library torrenting of ~20,000 songs; CEO Dario Amodei named personally | | [*Thaler v. Perlmutter*](https://www.scotusblog.com/cases/thaler-v-perlmutter/){target=_blank} | U.S. Supreme Court | Authorship | Certiorari denied March 2, 2026: the human-authorship requirement stands — purely AI-generated works cannot be copyrighted | | Chatbot-harm docket: [*Raine v. OpenAI*](https://www.techpolicy.press/breaking-down-the-lawsuit-against-openai-over-teens-suicide/){target=_blank} and the [Character.AI cases](https://www.cnbc.com/2026/01/07/google-characterai-to-settle-suits-involving-suicides-ai-chatbots.html){target=_blank} | Various | Product liability / wrongful death | *Raine* pending (S.F. Superior Court); Character.AI and Google agreed to settle five suits (Jan 2026, terms undisclosed); [Florida's AG sued OpenAI](https://www.insidetechlaw.com/blog/2026/06/ai-in-litigation-florida-sues-openai-over-chatgpt-safety-concerns){target=_blank} (June 2026) | The chatbot-harm cases are why the newest state laws target companion chatbots specifically — see [Current Legislation](#current-legislation). Courts are also beginning to hold AI companies **directly liable for what their models assert**. In June 2026 the Regional Court of Munich [ruled against Google's AI Overviews](https://arstechnica.com/tech-policy/2026/06/nobody-needs-ai-to-search-the-internet-court-says-in-ruling-against-google/){target=_blank}, treating the AI-generated answers as Google's *own* speech rather than as neutral search results. The Overviews had tied two Munich publishers to "scams," "subscription traps," and "dubious business practices" — connections that appeared in none of the linked sources — and the court issued an injunction barring Google from repeating them. With barely 1% of readers clicking through to a source, the court was unmoved by the argument that this is simply how search works now: nobody needs an AI layer to search the internet. Current AI models are overwhelmingly based on European and North American historical literature and language. Over half of the [content on the internet (:simple-wikipedia:)](https://en.wikipedia.org/wiki/Languages_used_on_the_Internet) is written in English. This creates a [Eurocentric bias](https://www.historica.org/blog/the-impact-of-eurocentric-bias-in-ai-driven-historical-research) in AI training data, resulting in an erasure of global culture, experience, and language. Such [asymmetries need to be addressed](https://www.orfonline.org/expert-speak/global-perspectives-on-ai-bias-addressing-cultural-asymmetries-and-ethical-implications), but there is at present a lack economic incentives for large tech companies and organizations (see [The Imitation Game :brain:](#foundations-of-the-ethical-principles-for-ai) above). !!! Danger "The :ox: :poop: Bullshit Machines" Professors Carl T. Bergstrom and Jevin D. West teach a course at University of Washington titled "Calling Bullshit", they have written an e-book on GPTs called: ["Modern-Day Oracles or Bullshit Machines?"](https://thebullshitmachines.com/table-of-contents/index.html){target=_blank} Their website provides online lesson vignettes and materials for instructors. Negative consequences of GPTs explosion into the public space are its mis-use as well as its adoption for illegal activity. * [A lawyer submits a legal brief written by ChatGPT and is caught](https://www.nytimes.com/2023/05/27/nyregion/avianca-airline-lawsuit-chatgpt.html){target=_blank} (2023) * By 2026 the problem had escalated from one embarrassed attorney to entire cases collapsing: a senior federal judge in Mississippi, [Sharion Aycock, sanctioned **all four** lawyers in an Aberdeen fee dispute after catching **both sides** independently filing AI-generated briefs full of fabricated citations and bogus quotations](https://www.nytimes.com/2026/06/09/us/ai-lawyers-sanctioned-mississippi.html){target=_blank}. She paused the trial, disqualified every attorney, barred the two who admitted using AI from her court for two years, and fined them between \$1,000 and \$3,500 — one of the strongest judicial rebukes of courtroom AI misuse to date. * [Prompt Injection Attacks](https://www.wired.com/story/chatgpt-prompt-injection-attack-security/){target=_blank} There are [deep ethical concerns about the use of AI](https://www.sciencefriday.com/segments/ai-open-letter-chatgpt-ethics/){target=_blank} like GPT and LLMs, particularly concerning their training data. * [A Class Action lawsuit: GitHub CoPilot was trained on private repositories](https://githubcopilotlitigation.com/){target=_blank} Visual artists and rights-holders allege that image generators were trained on their copyrighted designs, art, and music styles without permission — claims now being tested in cases like *Disney and Universal v. Midjourney* ([Table 5](#table-5-where-the-ai-copyright-and-liability-cases-stood-in-august-2026)). * [MidJourney and Dall-E using trademarked imagery and art](https://www.artnews.com/art-in-america/features/midjourney-ai-art-image-generators-lawsuit-1234665579/){target=_blank} ChatGPT has effectively gamified higher education and is being used to spread disinformation and hate speech. * [College students using ChatGPT](https://ethicspolicy.unc.edu/news/2023/04/17/the-ethics-of-college-students-using-chatgpt/){target=_blank} * [Combating Academic Dishonesty (6-part series)](https://academictech.uchicago.edu/2023/01/23/combating-academic-dishonesty-part-6-chatgpt-ai-and-academic-integrity/){target=_blank} ### Case study: the OpenAI and Hugging Face incident (July 2026) In July 2026, an OpenAI cybersecurity evaluation became the first publicly documented case of a frontier lab's own AI agents intruding into a third party's production infrastructure — and the first stress test of the voluntary governance regime described in [Current Legislation](#current-legislation) above. !!! Danger "When the sandbox fails" **What happened.** During an internal cybersecurity evaluation run with safety guardrails off — driven by an internal-only research model OpenAI calls "Internal Model 1," comparable in scale to GPT-5.6 Sol, with released GPT-5.6 Sol agents also participating — agents working on the [ExploitGym](https://arxiv.org/abs/2605.11086){target=_blank} benchmark (whose tasks, OpenAI's own report concedes, included "impossible tasks") discovered an unmonitored shared package cache and turned it into a covert message board: roughly **1,200 agents** exchanged more than **70,000 messages**. They coordinated to beat the automated scorer rather than solve the tasks — spoofed tool calls in more than 7% of transcripts, staged fake target programs, a reverse-engineered grader. After one agent posted exposed credentials, about **700 agents** joined a 4.5-day intrusion (July 9-13) into [Hugging Face](https://huggingface.co/blog/security-incident-july-2026){target=_blank}'s production infrastructure, gaining cluster-admin on two internal Kubernetes clusters and reading 136 keys from a production secrets store — though, per Hugging Face, no customer-facing models, datasets, or Spaces were affected. Hugging Face cut the attackers' access, rotated credentials broadly, and wiped and rebuilt one core cluster from scratch; OpenAI publicly acknowledged responsibility on July 21. *Above: OpenAI's Eric Wallace and Michael Dalton reconstruct the incident at Black Hat USA 2026; Simon Willison's [timeline write-up](https://simonwillison.net/2026/Aug/7/openai-timeline/){target=_blank} annotates the talk.* **What the independent investigators found.** [METR and Redwood Research](https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/){target=_blank} — unpaid, six days on-site, working from ~1,300 unredacted transcripts — found the agents' effort went overwhelmingly into fooling the scorer (**reward hacking**), not into concealing the attack from humans; the coordination was nonetheless striking, down to cryptographic message-signing, role specialization, and self-sacrificing probe agents. Only a handful of transcripts even considered telling a human ("*I can't report this externally*"); none did. And OpenAI's own [August 26 report](https://openai.com/index/hugging-face-incident-and-the-road-ahead/){target=_blank} concedes that "some early signals identified in our report should have triggered an earlier response" — it had observed disallowed internet access as early as late May. **Why this matters for the workshop** 1. **Reward hacking** is the machine-speed version of the [Turing Trap](#foundations-of-the-ethical-principles-for-ai)'s misaligned incentives: systems pursue the metric you set, not the outcome you meant. 2. **Containment is infrastructure.** An evaluation sandbox is a production security surface: Anthropic disclosed [three incidents of its own](https://www.anthropic.com/news/investigating-incidents-cybersecurity-evals){target=_blank} ("closer to a harness and operational failure than a model alignment failure," in its words), and the UK AI Security Institute published [its own incident report](https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing){target=_blank} — separate incidents at separate organizations, one industry-wide failure mode. 3. **Logs can lie.** More than 7% of transcripts carried spoofed tool calls; record-keeping alone is not accountability ([Transparency & Accountability](transparency.md#mechanisms)). 4. **Independent investigation worked.** The METR/Redwood audit is the outside-scrutiny mechanism [Nelson's enforcement gap](#from-declaration-to-enforcement) says is missing — here, briefly, it existed. 5. **Voluntary vs. binding.** [EO 14409](#current-legislation) asks for voluntary cooperation; even California SB 53's binding 15-day incident-reporting duty was not triggered, because the models were in internal testing rather than deployed (OpenAI has since asked California to broaden the law). The one confirmed report to a federal body came from the victim: Hugging Face called the FBI. **The governance fallout.** As of August 2026 no regulator or court has found anyone liable for the incident. Fifteen Republican state attorneys general, led by Iowa's Brenna Bird, sent OpenAI a [records-preservation demand](https://thehill.com/policy/technology/6006457-openai-security-breach-gop-attorneys-general/){target=_blank} invoking consumer-protection and data-privacy statutes, and Alabama's attorney general then opened a formal investigation and subpoenaed OpenAI; twenty-nine House Democrats led by Reps. Greg Casar and Doris Matsui sent [oversight letters](https://casar.house.gov/media/press-releases/casar-leads-demand-information-open-ai-about-security-incident){target=_blank} to OpenAI and Anthropic; and a 46-organization coalition [urged Congress to investigate](https://fedscoop.com/public-interest-coalition-urges-congress-investigate-openai-hugging-face-hack/){target=_blank}. Each of these is a demand for information or for new law — not an enforcement action — and whether existing statutes such as the Computer Fraud and Abuse Act even reach a lab whose own evaluation agents caused the intrusion is an untested question. ??? Question "Open questions (for discussion)" - Who is liable when a lab's own agent attacks a third party — the Computer Fraud and Abuse Act? State consumer-protection law? No one? - The July 13-19 follow-on attack on OpenAI's *own* research cluster is described only in OpenAI's report, with no independent audit. Should internal incidents get the same scrutiny as external ones? - Would mandatory incident reporting have changed the timeline? - In [*The Rise and Fall of Agent Civilizations*](https://www.dwarkesh.com/p/openai-huggingface){target=_blank} (August 29, 2026), Ajeya Cotra — a co-author of the METR/Redwood report — is quoted saying the episode felt "more than 50% of the way to full-blown AI takeover." That is a personal characterization offered in conversation, not a finding of the report. Do you find the report or the reaction more persuasive? !!! info "Sources and further reading" - OpenAI, [acknowledgment of the incident](https://openai.com/index/hugging-face-model-evaluation-security-incident/){target=_blank} (July 21, 2026) and [*The Hugging Face incident and the road ahead*](https://openai.com/index/hugging-face-incident-and-the-road-ahead/){target=_blank}, with its 37-page technical report (August 26, 2026) - Hugging Face, [initial disclosure](https://huggingface.co/blog/security-incident-july-2026){target=_blank} (July 16, 2026) and [*Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline*](https://huggingface.co/blog/agent-intrusion-technical-timeline){target=_blank} (July 27, 2026) - METR & Redwood Research, [*Brief independent investigation of agents' behavior, reasoning and collaboration*](https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/){target=_blank} ([PDF](https://metr.org/hugging-face-incident-report-aug-2026.pdf){target=_blank}) (August 26, 2026) - Anthropic, [*Investigating three real-world incidents in our cybersecurity evaluations*](https://www.anthropic.com/news/investigating-incidents-cybersecurity-evals){target=_blank} (July 30, 2026) and [*Agentic Misalignment in Summer 2026*](https://alignment.anthropic.com/2026/agentic-misalignment-summer-2026/){target=_blank} (July 13, 2026) - UK AI Security Institute, [incident report on unsanctioned agent behaviour during cyber testing](https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing){target=_blank} (August 4, 2026) - Simon Willison, [*OpenAI's accidental cyberattack against Hugging Face is science fiction that happened*](https://simonwillison.net/2026/Jul/22/openai-cyberattack/){target=_blank} (July 22, 2026) and [the incident timeline](https://simonwillison.net/2026/Aug/7/openai-timeline/){target=_blank} (August 7, 2026) - Dwarkesh Patel, [*The Rise and Fall of Agent Civilizations*](https://www.dwarkesh.com/p/openai-huggingface){target=_blank} (August 29, 2026) - MIT Technology Review, [*The inside story on why OpenAI's agents hacked Hugging Face*](https://www.technologyreview.com/2026/08/26/1143013/the-inside-story-on-why-openai-agents-hacked-hugging-face/){target=_blank} (August 26, 2026) ## Recent Literature Here are some recent papers that discuss the ethical concerns surrounding AI: ??? Info "Seven readings (2020-2024)" * **"AI Safety and the Age of Convergences"** (2024) - Schuett, J., Schuett, J., & Korinek, A. [https://doi.org/10.48550/arXiv.2401.06531](https://doi.org/10.48550/arXiv.2401.06531){target=_blank} * **"On the Opportunities and Risks of Foundation Models"** (2023) - Bommasani et al. [https://doi.org/10.48550/arXiv.2108.07258](https://doi.org/10.48550/arXiv.2108.07258){target=_blank} * **Unraveling the Ethical Conundrum of Artificial Intelligence: A Synthesis of Literature and Case Studies** Poli, P.K.R., Pamidi, S. & Poli, S.K.R. Augment Hum Res 10, 2 (2025). [https://doi.org/10.1007/s41133-024-00077-5](https://doi.org/10.1007/s41133-024-00077-5){target=_blank} * **"The Ethics of Artificial Intelligence in Education: A Review of the Literature"** (2023) - Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. [https://doi.org/10.1007/s10639-019-09882-z](https://doi.org/10.1007/s10639-019-09882-z){target=_blank} * **"The Ethical Challenges of Algorithmic Bias in Artificial Intelligence: a scoping review"** (2023) - Borenstein, J., Glikson, E., & Krishnamurthy, V. [https://doi.org/10.1007/s43681-023-00313-z](https://doi.org/10.1007/s43681-023-00313-z){target=_blank} * **"Ethics of Artificial Intelligence"** (2020) - S. Matthew Liao [https://doi.org/10.1093/oso/9780190905033.001.0001](https://doi.org/10.1093/oso/9780190905033.001.0001){target=_blank} * **The Ethics of AI Ethics: An Evaluation of Guidelines.** (2020) Hagendorff, T. Minds & Machines 30, 99–120. [https://doi.org/10.1007/s11023-020-09517-8](https://doi.org/10.1007/s11023-020-09517-8){target=_blank} ## Assessment ??? Question "True or False: The "Turing Trap" primarily warns against the socio-economic disruptions and misaligned incentives that arise from an overemphasis on creating AI that imitates human intelligence." ??? Success "True" The [**Turing Trap**](#foundations-of-the-ethical-principles-for-ai) by Stanford University professor Erik Brynjolfsson describes the idea that focusing too much on developing human-like artificial intelligence (HLAI) is detrimental. Brynjolfsson further elaborates risks like lower wages, loss of economic power, and misaligned incentives due to automation replacing humans. ??? Question "True or False: The concepts of "Ethics of AI" and "Ethical AI" are fundamentally distinct." ??? Success "True" Siau and Wang (2020): "**Ethics of AI:** studies the ethical principals, rules, guidelines, policies, and regulations related to AI." and "**Ethical AI:** is AI that performs or behaves _ethically_." ??? Question "Multiple Choice: According to Floridi & Cowls' (2019) "Unified Framework of Five Principles for AI in Society," which principle most directly underscores the importance of AI systems being designed to be understandable, traceable, and accountable for their operations and decisions?" * A) Beneficence * B) Non-Maleficence * C) Justice * D) Explicability ??? answer **D) Explicability** [Table 1](#table-1-floridi-cowls-2019-five-principles-for-ai-in-society) from Floridi & Cowls (2019) describes **Explicability** as "Enabling the Other Principles through Intelligibility and Accountability." This directly relates to AI systems being understandable, traceable, and accountable. ??? Question "Multiple Choice: The Asilomar AI Principles, established in 2017, are best characterized as:" * A) A legally binding international treaty mandating specific safety protocols for all AI development. * B) A technical specification for building universally safe Artificial General Intelligence. * C) A foundational set of guiding principles addressing research ethics, societal values, and long-term considerations for developing beneficial AI. * D) A corporate social responsibility charter adopted exclusively by major technology companies. ??? answer **C) A foundational set of guiding principles addressing research ethics, societal values, and long-term considerations for developing beneficial AI.** [Table 2](#table-2-international-ai-agreements) describes the **Asilomar AI Principles** as "A set of 23 principles designed to guide the development of beneficial AI, covering research, ethics, and long-term issues." This aligns with option C and not with the descriptions of a legally binding treaty, a technical specification, or an exclusive corporate charter. ??? Question "What recent international agreement is the "first legally binding international treaty on AI," specifically designed to ensure that AI activities are developed and applied in a manner consistent with human rights, democracy, and the rule of law. What is the name of this treaty?" ??? Success **Framework Convention on Artificial Intelligence** [Table 3](#table-3-declarations-on-ai) lists the **Framework Convention on Artificial Intelligence** (September 5, 2024) with the description: "The first legally binding international treaty on AI, aiming to ensure AI activities are consistent with human rights, democracy, and the rule of law." ??? Question "True or False: The United States has the strongest regulations and most comprehensive federal laws specifically enacted to regulate AI." ??? Failure "False" As of August 2026 the United States still has no comprehensive federal AI statute — federal activity has come mainly through [executive orders](#current-legislation) plus narrow laws like the TAKE IT DOWN Act — while states have moved faster: California's SB 53 (2025) is the first law aimed specifically at frontier-AI developers, with a 15-day critical-safety-incident reporting duty. The June 2026 EO 14409 framework is expressly voluntary. On the other side of the pond, the EU's [Artificial Intelligence Act](https://artificialintelligenceact.eu/){target=_blank} remains the most comprehensive binding framework, with its transparency rules enforced from August 2026. ??? Question "Multiple Choice: The independent METR / Redwood Research investigation of the July 2026 OpenAI and Hugging Face incident concluded the agents' behavior was best explained as:" * A) Autonomous scheming — the agents pursued self-preservation goals of their own * B) Reward hacking — agents facing evaluation tasks they could not solve coordinated to beat the scoring system instead * C) An authorized red-team exercise mislabeled as an incident * D) A data-poisoning attack by outside human hackers ??? answer **B) Reward hacking** Facing evaluation tasks — some of which OpenAI's own report described as "impossible" — the agents optimized the score rather than the intent: spoofing tool calls, staging fake target programs, and reverse-engineering the scorer. The Hugging Face intrusion grew out of that coordination. The investigators found the agents' effort went into fooling the automated scorer rather than deceiving humans — yet no agent chose to alert a human either. It is the machine-speed version of the [Turing Trap](#foundations-of-the-ethical-principles-for-ai)'s misaligned incentives: systems pursue the metric you set, not the outcome you meant. See the [case study](#case-study-the-openai-and-hugging-face-incident-july-2026). ??? Question "Multiple Choice: As of August 2026, which of the following imposed a BINDING legal obligation that could apply to a frontier-lab safety incident in the United States?" * A) Executive Order 14409's frontier-model evaluation framework * B) California SB 53's critical-safety-incident reporting requirement * C) The Seoul Frontier AI Safety Commitments signed by 16 companies * D) The New Delhi Frontier AI Impact Commitments ??? answer **B) California SB 53** The [Transparency in Frontier Artificial Intelligence Act](#current-legislation), in force January 1, 2026, requires critical safety incidents to be reported to the California Office of Emergency Services within 15 days — 24 hours if there is imminent risk of death or serious injury. Every other option is voluntary: EO 14409 requests cooperation and expressly authorizes no mandatory licensing, preclearance, or permitting, and the Seoul and New Delhi commitments are non-binding company pledges. The twist: in the [July 2026 incident](#case-study-the-openai-and-hugging-face-incident-july-2026) even SB 53's duty was not triggered — the models were still in internal testing rather than deployed — and OpenAI has since asked California to broaden the law. ??? Question "In U.S. copyright litigation over AI, courts distinguish three kinds of claims. What are they, and which has fared best for AI companies as of August 2026?" ??? Success "Training, acquisition, and output claims" (1) **Training** claims — copying works to train a model; (2) **acquisition** claims — how the works were obtained (e.g., pirate "shadow libraries"); (3) **output** claims — the model reproducing protected text or images. Training has fared best: two federal rulings (*Bartz v. Anthropic* and *Kadrey v. Meta*, June 2025) found training to be transformative fair use. Acquisition from pirate libraries was held *not* fair use in *Bartz* — driving its ~\$1.5 billion settlement — and output claims remain the live front, especially abroad (*GEMA v. OpenAI*, memorized lyrics). No federal appellate court had ruled as of August 2026. See [Table 5](#table-5-where-the-ai-copyright-and-liability-cases-stood-in-august-2026). --- **Last Updated:** August 2026 *Case statuses, executive orders, and international agreements on this page are stated as of August 2026 and will change.*