The Chronicle

President calls people who say 'Artificial Intelligence' instead of 'Super Intelligence' 'THE ENEMY!'

Nine days after Executive Order 14434, the President wrote that anyone using 'Artificial Intelligence' rather than 'Super Intelligence' is 'THE ENEMY!'. Axios reported that White House pages still used 'AI'.

What it asks to be measured. None.

Source: Axios (secondary) · Forbes (secondary). Checked Oct 09, 2026.

Limits: The post itself was not retrieved. Sources were found by web search on Oct 09, 2026 and could not be opened from our fetcher; the summary rests on the search results and their quoted text, not a full read.

Cite. Protocol STAMP, “President calls people who say 'Artificial Intelligence' instead of 'Super Intelligence' 'THE ENEMY!',” The Chronicle, entry us-fed-trump-si-enemy-2026-10-08, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#us-fed-trump-si-enemy-2026-10-08.

US Defense Secretary orders an Autonomous Warfare Command

A memo directs the department to begin building a four-star combatant command with service-like authorities to scale autonomous and robotic capabilities across the joint force, to stand up by Oct 01, 2027.

What it asks to be measured. None.

Source: Defense One (secondary) · Inside Defense (memo text) (secondary). Checked Oct 09, 2026.

Limits: The memo was not retrieved. Sources were found by web search on Oct 09, 2026 and could not be opened from our fetcher; the summary rests on the search results and their quoted text, not a full read.

Cite. Protocol STAMP, “US Defense Secretary orders an Autonomous Warfare Command,” The Chronicle, entry us-fed-autonomous-warfare-command-2026-09-30, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#us-fed-autonomous-warfare-command-2026-09-30.

D.C. Circuit upholds the Pentagon's designation of Anthropic, 2 to 1

In Anthropic's separate challenge to the designation, a D.C. Circuit panel upheld it by two votes to one. The majority found the department had enough evidence to conclude that Claude's built-in restrictions and the unresolved contract dispute could make it unreliable for military operations; the dissent read the statute differently. Anthropic said it disagreed with the decision.

What it asks to be measured. None.

Source: ABC News (secondary) · Let's Data Science (secondary). Checked Oct 09, 2026.

Limits: Disclosure: Anthropic, a party here, makes Claude, a tool this site uses; this entry was drafted and gated by a Claude model. The opinion was not retrieved, and its citation was not reported in the coverage found. Sources were found by web search on Oct 09, 2026 and could not be opened from our fetcher; the summary rests on the search results and their quoted text, not a full read.

Cite. Protocol STAMP, “D.C. Circuit upholds the Pentagon's designation of Anthropic, 2 to 1,” The Chronicle, entry ct-anthropic-dccir-2026-09-25, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#ct-anthropic-dccir-2026-09-25.

State Department tells staff to say 'SI' for AI

Following the President's renaming of AI as 'Super Intelligence' at the UN on Sep 22, 2026, an internal State Department email told staff to use 'SI' in all remarks, positions and press elements, as reported by the Associated Press. US representatives used the term at the UN Security Council.

What it asks to be measured. None.

Source: Associated Press via Yahoo News (secondary) · Euronews (secondary). Checked Oct 09, 2026.

Limits: The email was not retrieved; the date is that of the first report found, not the email's own. See the Docket entry on Executive Order 14434. Sources were found by web search on Oct 09, 2026 and could not be opened from our fetcher; the summary rests on the search results and their quoted text, not a full read.

Cite. Protocol STAMP, “State Department tells staff to say 'SI' for AI,” The Chronicle, entry us-fed-state-si-directive-2026-09-24, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#us-fed-state-si-directive-2026-09-24.

UN expert group on lethal autonomous weapons ends its last session without a binding instrument

The CCW Group of Governmental Experts held its last session under its mandate from Aug 31, 2026. It ran into the early hours of Sep 05, 2026 and adopted a final report, without a legally binding agreement; Human Rights Watch called the text significantly watered down. The report goes to the Seventh CCW Review Conference in November 2026. 76 states now support negotiating a legally binding instrument.

What it asks to be measured. None.

Source: UN Office for Disarmament Affairs · Davis Vanguard (secondary) · Automated Research (secondary) · Human Rights Watch (secondary). Checked Oct 10, 2026.

Limits: Corrected Oct 10, 2026: the entry first said the session ended Sep 04, 2026 with a draft report. The UN's final report was not retrieved; the timing, the adoption and the 76 states come from Human Rights Watch (Sep 07, 2026), read in full from an archived copy. Whether it was adopted by consensus is not confirmed.

Cite. Protocol STAMP, “UN expert group on lethal autonomous weapons ends its last session without a binding instrument,” The Chronicle, entry un-ccw-gge-laws-ends-2026-09-04, last checked Oct 10, 2026, https://protocolstamp.com/chronicle/#un-ccw-gge-laws-ends-2026-09-04.

Anthropic v. Department of War: court vacates the supply-chain designation

Granting summary judgment, Judge Rita Lin held that the designation was retaliation against Anthropic's speech in violation of the First Amendment, denied due process under the Fifth Amendment, and was arbitrary and capricious; she vacated it. The order does not require the government to buy Anthropic's products. A separate challenge in the D.C. Circuit was still pending.

What it asks to be measured. None.

Source: CNBC (secondary) · CBS News (secondary) · The Hill (secondary). Checked Oct 09, 2026.

Limits: Disclosure: Anthropic, a party here, makes Claude, a tool this site uses; this entry was drafted and gated by a Claude model. The judgment was not retrieved; reports of the evening of Aug 27, 2026 are dated Aug 28, 2026. Sources were found by web search on Oct 09, 2026 and could not be opened from our fetcher; the summary rests on the search results and their quoted text, not a full read.

Cite. Protocol STAMP, “Anthropic v. Department of War: court vacates the supply-chain designation,” The Chronicle, entry ct-anthropic-dow-ndcal-2026-08-27, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#ct-anthropic-dow-ndcal-2026-08-27.

Court grants final approval of the USD 1.5 billion Bartz v. Anthropic settlement

The Northern District of California granted final approval of the class settlement and entered judgment, about USD 3,000 per eligible work. Anthropic must destroy copies obtained from pirate libraries; the settlement covers past conduct only.

What it asks to be measured. None.

Source: The Authors Guild (supported the class; secondary) (secondary) · CourtListener (Free Law Project) docket (secondary). Checked Oct 08, 2026.

Limits: The final-approval order itself was not retrieved; date (Jul 20, 2026) and terms from the Authors Guild and news reports. Reports name Judge Araceli Martinez-Olguin as approving after Judge Alsup's retirement.

See explainer: AI and copyright: rulings so far.

Cite. Protocol STAMP, “Court grants final approval of the USD 1.5 billion Bartz v. Anthropic settlement,” The Chronicle, entry ct-bartz-final-approval-2026-07-20, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-bartz-final-approval-2026-07-20.

South Africa withdraws its draft National AI Policy after fictitious references are found

The Minister of Communications and Digital Technologies withdrew the draft National AI Policy, approved by Cabinet on Mar 25, 2026 and gazetted for comment, after internal checks confirmed that its reference list contained fictitious sources, which he linked to unverified AI-generated citations. Cabinet approved the withdrawal; a revised policy is to return for comment.

What it asks to be measured. None directly; a case of unverified AI-generated citations in an official policy draft.

Source: SAnews (South African Government News Agency) · SAnews (South African Government News Agency). Checked Oct 09, 2026.

Limits: The link to AI-generated citations is from the Minister's budget vote speech as reported; the gazetted withdrawal notice was not read.

Cite. Protocol STAMP, “South Africa withdraws its draft National AI Policy after fictitious references are found,” The Chronicle, entry za-ai-policy-withdrawn-2026-04-26, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#za-ai-policy-withdrawn-2026-04-26.

Anthropic v. Department of War: preliminary injunction against the supply-chain designation

In the Northern District of California, where Anthropic sued on Mar 09, 2026, Judge Rita Lin preliminarily enjoined the designation and the directive that agencies stop using Anthropic's services, pending a decision on the merits. The order was stayed for seven days to allow an appeal.

What it asks to be measured. None.

Source: NPR (secondary) · CNBC (secondary) · FedScoop (secondary). Checked Oct 09, 2026.

Limits: Disclosure: Anthropic, a party here, makes Claude, a tool this site uses; this entry was drafted and gated by a Claude model. The order was not retrieved. Sources were found by web search on Oct 09, 2026 and could not be opened from our fetcher; the summary rests on the search results and their quoted text, not a full read.

Cite. Protocol STAMP, “Anthropic v. Department of War: preliminary injunction against the supply-chain designation,” The Chronicle, entry ct-anthropic-dow-pi-2026-03-26, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#ct-anthropic-dow-pi-2026-03-26.

The Pentagon designates Anthropic a supply-chain risk

The President directed federal agencies to stop using Anthropic's models, and the Secretary of Defense designated the company a supply-chain risk, the first US company so designated; formal letters are dated Mar 03, 2026. Anthropic had declined contract terms allowing 'any lawful use' of its models, which would have removed its limits on fully autonomous weapons and domestic mass surveillance.

What it asks to be measured. None.

Source: NPR (secondary) · Mayer Brown (secondary) · Congressional Research Service (secondary). Checked Oct 09, 2026.

Limits: Disclosure: Anthropic, a party here, makes Claude, a tool this site uses; this entry was drafted and gated by a Claude model and whose models this program measures. The designation letters were not retrieved. Sources were found by web search on Oct 09, 2026 and could not be opened from our fetcher; the summary rests on the search results and their quoted text, not a full read.

Cite. Protocol STAMP, “The Pentagon designates Anthropic a supply-chain risk,” The Chronicle, entry us-fed-anthropic-supply-chain-2026-02-27, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#us-fed-anthropic-supply-chain-2026-02-27.

UN General Assembly resolution 80/57 on lethal autonomous weapons systems

The General Assembly's third consecutive resolution on lethal autonomous weapons systems. It urges the CCW expert group to complete elements for an instrument, with a view to negotiation. The First Committee approved it on Nov 06, 2025 by 156 to 5, with 8 abstentions; the Assembly adopted it on Dec 01, 2025 by 164 to 6, with 7 abstentions.

What it asks to be measured. None.

Source: Stop Killer Robots (campaign; secondary) (secondary) · Digital Watch (secondary). Checked Oct 09, 2026.

Limits: Corrected Oct 09, 2026: the entry first gave 156, the First Committee count, as the Assembly's. Vote figures come from the UN press release GA/12736 as quoted in search results and the UN Digital Library record 4095989; neither could be opened from our fetcher, so the entry stays Likely.

Cite. Protocol STAMP, “UN General Assembly resolution 80/57 on lethal autonomous weapons systems,” The Chronicle, entry un-ga-res-80-57-2025-12-01, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#un-ga-res-80-57-2025-12-01.

GEMA v OpenAI: Munich Regional Court finds song lyrics memorised by ChatGPT infringe

The 42nd Civil Chamber of the Landgericht München I (case 42 O 14139/24) held that OpenAI's models memorised and reproduced protected German song lyrics, which infringed copyright, and that the text-and-data-mining exception did not cover it.

What it asks to be measured. None directly; memorisation of training data as a legal test.

Source: Legal Tribune Online (secondary) (secondary). Checked Oct 08, 2026.

Limits: Not final. The court's own press release was not retrieved; summary from legal press citing the judgment (Az. 42 O 14139/24).

See explainer: AI and copyright: rulings so far.

Cite. Protocol STAMP, “GEMA v OpenAI: Munich Regional Court finds song lyrics memorised by ChatGPT infringe,” The Chronicle, entry ct-gema-openai-2025-11-11, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-gema-openai-2025-11-11.

Getty Images v Stability AI: UK High Court judgment

The High Court (Joanna Smith J) largely dismissed Getty's remaining claims, holding that the Stable Diffusion model weights are not an 'infringing copy' because they do not store the training images; Getty succeeded in part on trade marks for some older model versions.

What it asks to be measured. None.

Source: Courts and Tribunals Judiciary (England and Wales) · Courts and Tribunals Judiciary (judgment PDF). Checked Oct 08, 2026.

See explainer: AI and copyright: rulings so far.

Cite. Protocol STAMP, “Getty Images v Stability AI: UK High Court judgment,” The Chronicle, entry ct-getty-stability-ukhc-2025-11-04, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-getty-stability-ukhc-2025-11-04.

Microsoft disables some cloud and AI services to a unit of Israel's Ministry of Defence

After reporting by The Guardian, +972 Magazine and Local Call that Unit 8200 stored intercepted phone-call data on Azure, Microsoft's review found evidence supporting elements of the reporting, including use of Azure storage in the Netherlands and of AI services, and it disabled specific services to a unit of the ministry. Israel's statements on its use of Microsoft services varied.

What it asks to be measured. None.

Source: NBC News (secondary) · Al Jazeera (secondary). Checked Oct 09, 2026.

Limits: Microsoft's own statement was not retrieved. Sources were found by web search on Oct 09, 2026 and could not be opened from our fetcher; the summary rests on the search results and their quoted text, not a full read.

Cite. Protocol STAMP, “Microsoft disables some cloud and AI services to a unit of Israel's Ministry of Defence,” The Chronicle, entry ms-microsoft-imod-azure-2025-09-25, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#ms-microsoft-imod-azure-2025-09-25.

Bartz v. Anthropic: training on lawfully acquired books is fair use; pirated library is not

Judge William Alsup (N.D. Cal.) held that training Claude on purchased and scanned books was fair use, but that building a central library from pirated copies was not. The case later settled as a class action for USD 1.5 billion, which the court preliminarily approved in September 2025.

What it asks to be measured. None.

Source: US District Court, N.D. Cal. (Order on Fair Use, via RECAP) · CourtListener (Free Law Project) docket (secondary). Checked Oct 08, 2026.

Limits: Settlement amount and approval date from news reports; confirm against the court's approval order.

See explainer: AI and copyright: rulings so far.

Cite. Protocol STAMP, “Bartz v. Anthropic: training on lawfully acquired books is fair use; pirated library is not,” The Chronicle, entry ct-bartz-anthropic-2025-06-23, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-bartz-anthropic-2025-06-23.

Pope Leo XIV ties his papal name to AI in his first address to the College of Cardinals

Explaining his choice of name, the new pope recalled Leo XIII's response to the first industrial revolution and said the Church offers its social teaching in response to developments in artificial intelligence that challenge human dignity, justice and labour.

What it asks to be measured. None stated.

“another industrial revolution and to developments in the field of artificial intelligence”Pope Leo XIV, official Vatican transcript

Source: The Holy See (vatican.va). Checked Oct 08, 2026.

See explainer: Landmark speeches on AI.

Cite. Protocol STAMP, “Pope Leo XIV ties his papal name to AI in his first address to the College of Cardinals,” The Chronicle, entry va-pope-leo-cardinals-ai-2025-05-10, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#va-pope-leo-cardinals-ai-2025-05-10.

US Copyright Office pre-publication report on generative AI training (Part 3)

A pre-publication version said some training uses may be fair use and others not, depending on factors including whether outputs compete with the works trained on. The Register of Copyrights, Shira Perlmutter, was dismissed the next day, a removal she challenged in court.

What it asks to be measured. None.

Source: US Copyright Office. Checked Oct 08, 2026.

Limits: The dismissal and litigation are from news reporting, not the report.

Cite. Protocol STAMP, “US Copyright Office pre-publication report on generative AI training (Part 3),” The Chronicle, entry us-fed-copyright-part3-2025-05-09, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#us-fed-copyright-part3-2025-05-09.

US Court of Appeals (D.C. Cir.): work generated autonomously by AI is not copyrightable (Thaler v. Perlmutter)

Affirmed the 2023 district-court ruling that the Copyright Act requires human authorship, upholding the Copyright Office's refusal to register a work listing the 'Creativity Machine' as author.

What it asks to be measured. None.

Source: US Court of Appeals for the D.C. Circuit. Checked Oct 08, 2026.

See explainer: AI and copyright: rulings so far.

Cite. Protocol STAMP, “US Court of Appeals (D.C. Cir.): work generated autonomously by AI is not copyrightable (Thaler v. Perlmutter),” The Chronicle, entry ct-thaler-perlmutter-2025-03-18, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-thaler-perlmutter-2025-03-18.

Thomson Reuters v. Ross Intelligence: fair-use defence rejected for AI training on Westlaw headnotes

Judge Stephanos Bibas (D. Del., sitting by designation) granted partial summary judgment to Thomson Reuters, holding that Ross's copying of Westlaw headnotes to build a competing (non-generative) AI legal-research tool was not fair use. Interlocutory appeal to the Third Circuit followed.

What it asks to be measured. None.

Source: US District Court for the District of Delaware. Checked Oct 08, 2026.

See explainer: AI and copyright: rulings so far.

Cite. Protocol STAMP, “Thomson Reuters v. Ross Intelligence: fair-use defence rejected for AI training on Westlaw headnotes,” The Chronicle, entry ct-thomson-reuters-ross-2025-02-11, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-thomson-reuters-ross-2025-02-11.

Vice President Vance at the Paris AI Action Summit argues against excessive AI regulation

At the AI Action Summit in Paris, the Vice President said the United States would favour pro-growth AI policies and argued that excessive regulation could harm the AI industry. The US did not sign the summit statement.

What it asks to be measured. None stated.

“could kill a transformative industry just as it's taking off”JD Vance, Vice President of the United States, transcript, American Presidency Project

Source: The American Presidency Project, UC Santa Barbara (transcript archive) (secondary). Checked Oct 08, 2026.

Limits: No White House copy was found (404); the transcript is from a university archive. Non-signature of the statement is from the Paris statement entry's sources.

See explainer: Landmark speeches on AI.

Cite. Protocol STAMP, “Vice President Vance at the Paris AI Action Summit argues against excessive AI regulation,” The Chronicle, entry us-fed-vance-paris-2025-02-11, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#us-fed-vance-paris-2025-02-11.

Italy's Garante orders DeepSeek blocked

Ordered an urgent limitation on DeepSeek's processing of Italian users' data after the companies said they were not subject to EU law; the app was removed from Italian app stores. Other states restricted DeepSeek on government devices in the following weeks.

What it asks to be measured. None directly.

Source: Garante per la protezione dei dati personali. Checked Oct 08, 2026.

See explainer: Regulators' actions against AI uses.

Cite. Protocol STAMP, “Italy's Garante orders DeepSeek blocked,” The Chronicle, entry en-it-garante-deepseek-2025-01-30, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#en-it-garante-deepseek-2025-01-30.

US Copyright Office report on the copyrightability of AI outputs (Part 2)

Concluded that existing law suffices: outputs are protectable only where a human author has determined sufficient expressive elements; prompts alone generally do not give enough control.

What it asks to be measured. None.

Source: US Copyright Office. Checked Oct 08, 2026.

Cite. Protocol STAMP, “US Copyright Office report on the copyrightability of AI outputs (Part 2),” The Chronicle, entry us-fed-copyright-part2-2025-01-29, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#us-fed-copyright-part2-2025-01-29.

Stargate Project announced

OpenAI announced Stargate, a new company intending to invest USD 500 billion over four years in AI infrastructure for OpenAI in the United States, deploying USD 100 billion immediately. Initial equity funders: SoftBank, OpenAI, Oracle and MGX.

What it asks to be measured. None.

Source: OpenAI. Checked Oct 08, 2026.

Cite. Protocol STAMP, “Stargate Project announced,” The Chronicle, entry ms-stargate-2025-01-21, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-stargate-2025-01-21.

DeepSeek releases R1 reasoning model with open weights

The Chinese lab DeepSeek released DeepSeek-R1 under an MIT licence, reporting reasoning-benchmark performance comparable to OpenAI's o1. The release moved markets and prompted several governments to restrict the app on official devices.

What it asks to be measured. Self-reported benchmark results (e.g. AIME 2024, MATH-500); open weights allow independent evaluation.

Source: arXiv (DeepSeek-AI) · Hugging Face (DeepSeek-AI model card). Checked Oct 08, 2026.

See explainer: Technology milestones that shaped the rules.

Cite. Protocol STAMP, “DeepSeek releases R1 reasoning model with open weights,” The Chronicle, entry ms-deepseek-r1-2025-01-20, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-deepseek-r1-2025-01-20.

Italy's Garante fines OpenAI EUR 15 million over ChatGPT

Concluded the 2023 investigation with a EUR 15 million fine and an order for a six-month public information campaign about how ChatGPT collects data for training. OpenAI said it would appeal.

What it asks to be measured. Transparency to users and non-users about training-data processing.

Source: Garante per la protezione dei dati personali. Checked Oct 08, 2026.

See explainer: Regulators' actions against AI uses.

Cite. Protocol STAMP, “Italy's Garante fines OpenAI EUR 15 million over ChatGPT,” The Chronicle, entry en-it-garante-openai-fine-2024-12-20, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#en-it-garante-openai-fine-2024-12-20.

Biden and Xi affirm human control over the decision to use nuclear weapons

Meeting in Lima at APEC, the US and Chinese presidents affirmed the need to keep human control over decisions to use nuclear weapons, and to develop military AI prudently and responsibly. It was the first such statement by China; it is not binding and set no follow-up.

What it asks to be measured. None.

Source: The White House (archived readout) · NPR (secondary). Checked Oct 09, 2026.

Limits: Sources were found by web search on Oct 09, 2026 and could not be opened from our fetcher; the summary rests on the search results and their quoted text, not a full read.

Cite. Protocol STAMP, “Biden and Xi affirm human control over the decision to use nuclear weapons,” The Chronicle, entry intl-biden-xi-nuclear-ai-2024-11-16, last checked Oct 09, 2026, https://protocolstamp.com/chronicle/#intl-biden-xi-nuclear-ai-2024-11-16.

Nobel Prize in Chemistry 2024 to David Baker, Demis Hassabis and John Jumper

Half to David Baker for computational protein design and half jointly to Demis Hassabis and John Jumper for protein structure prediction (AlphaFold).

What it asks to be measured. None.

Source: The Nobel Prize (Royal Swedish Academy of Sciences). Checked Oct 08, 2026.

See explainer: Technology milestones that shaped the rules.

Cite. Protocol STAMP, “Nobel Prize in Chemistry 2024 to David Baker, Demis Hassabis and John Jumper,” The Chronicle, entry ms-nobel-chemistry-2024-10-09, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-nobel-chemistry-2024-10-09.

Nobel Prize in Physics 2024 to John Hopfield and Geoffrey Hinton

Awarded 'for foundational discoveries and inventions that enable machine learning with artificial neural networks'.

What it asks to be measured. None.

Source: The Nobel Prize (Royal Swedish Academy of Sciences). Checked Oct 08, 2026.

See explainer: Technology milestones that shaped the rules.

Cite. Protocol STAMP, “Nobel Prize in Physics 2024 to John Hopfield and Geoffrey Hinton,” The Chronicle, entry ms-nobel-physics-2024-10-08, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-nobel-physics-2024-10-08.

Dutch Data Protection Authority fines Clearview AI EUR 30.5 million

The Autoriteit Persoonsgegevens fined Clearview EUR 30.5 million for building an illegal database of facial images, with further penalty payments of up to EUR 5.1 million if violations continue.

What it asks to be measured. None directly.

Source: Autoriteit Persoonsgegevens (in Dutch). Checked Oct 08, 2026.

See explainer: Regulators' actions against AI uses.

Cite. Protocol STAMP, “Dutch Data Protection Authority fines Clearview AI EUR 30.5 million,” The Chronicle, entry en-nl-ap-clearview-2024-09-03, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#en-nl-ap-clearview-2024-09-03.

Tennessee ELVIS Act protects voice against AI cloning

The Ensuring Likeness Voice and Image Security (ELVIS) Act updated Tennessee's Protection of Personal Rights law to protect songwriters', performers' and others' voices from misuse of AI, including through tools and services that enable unauthorised impersonation.

What it asks to be measured. None.

Source: Office of the Governor of Tennessee. Checked Oct 08, 2026.

Limits: The Jul 01, 2024 effective date is from secondary reporting, not the Governor's release.

Cite. Protocol STAMP, “Tennessee ELVIS Act protects voice against AI cloning,” The Chronicle, entry us-tn-elvis-act-2024-03-21, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#us-tn-elvis-act-2024-03-21.

Brazil's Superior Electoral Court regulates AI in election campaigns

On Feb 27, 2024 the court amended Resolution 23.610/2019 on electoral propaganda to ban deepfakes and require notice of AI use in campaign content, and to restrict the use of bots to contact voters, for the October 2024 municipal elections.

What it asks to be measured. Labelling of synthetic campaign content.

Source: Tribunal Superior Eleitoral. Checked Oct 08, 2026.

Cite. Protocol STAMP, “Brazil's Superior Electoral Court regulates AI in election campaigns,” The Chronicle, entry br-tse-elections-ai-2024-02-27, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#br-tse-elections-ai-2024-02-27.

Google pauses Gemini image generation of people after inaccurate historical images

Google said Gemini's image generation 'got it wrong', producing inaccurate or offensive depictions, and paused generating images of people while it worked on a fix.

What it asks to be measured. Pre-release testing gaps acknowledged by the developer.

Source: Google. Checked Oct 08, 2026.

Cite. Protocol STAMP, “Google pauses Gemini image generation of people after inaccurate historical images,” The Chronicle, entry ms-gemini-image-pause-2024-02-23, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-gemini-image-pause-2024-02-23.

Moffatt v. Air Canada: airline liable for its chatbot's wrong answer

British Columbia's Civil Resolution Tribunal rejected Air Canada's argument that its website chatbot was a separate entity responsible for its own actions, and ordered the airline to honour the bereavement-fare information the chatbot gave.

What it asks to be measured. None directly; companies answer for what their chatbots tell customers.

Source: Civil Resolution Tribunal of British Columbia (2024 BCCRT 149). Checked Oct 08, 2026.

Cite. Protocol STAMP, “Moffatt v. Air Canada: airline liable for its chatbot's wrong answer,” The Chronicle, entry ct-moffatt-air-canada-2024-02-14, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-moffatt-air-canada-2024-02-14.

USPTO inventorship guidance for AI-assisted inventions

Clarified that AI-assisted inventions are not categorically unpatentable, but each named inventor must be a natural person who made a significant contribution, applying the Pannu factors.

What it asks to be measured. None.

Source: Federal Register. Checked Oct 08, 2026.

Cite. Protocol STAMP, “USPTO inventorship guidance for AI-assisted inventions,” The Chronicle, entry us-fed-uspto-inventorship-2024-02-13, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#us-fed-uspto-inventorship-2024-02-13.

The New York Times sues Microsoft and OpenAI for copyright infringement

Complaint filed in the Southern District of New York alleging that millions of Times articles were used without permission to train GPT models and that the models can reproduce Times content nearly verbatim.

What it asks to be measured. Memorisation and regurgitation of training data (examples of near-verbatim output in the complaint).

Source: The New York Times Company (complaint PDF). Checked Oct 08, 2026.

Cite. Protocol STAMP, “The New York Times sues Microsoft and OpenAI for copyright infringement,” The Chronicle, entry ct-nyt-openai-complaint-2023-12-27, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-nyt-openai-complaint-2023-12-27.

UK Supreme Court: an AI system cannot be named as an inventor (Thaler v Comptroller-General)

Held that under the Patents Act 1977 an inventor must be a natural person, so applications naming the AI system DABUS as inventor were properly refused. Part of a worldwide test-case series by Stephen Thaler.

What it asks to be measured. None.

Source: The Supreme Court of the United Kingdom. Checked Oct 08, 2026.

See explainer: AI and copyright: rulings so far.

Cite. Protocol STAMP, “UK Supreme Court: an AI system cannot be named as an inventor (Thaler v Comptroller-General),” The Chronicle, entry ct-thaler-dabus-uksc-2023-12-20, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-thaler-dabus-uksc-2023-12-20.

OpenAI board removes Sam Altman as CEO; reinstated days later

OpenAI's nonprofit board announced that Altman was departing after it concluded he was 'not consistently candid'. After staff and investor pressure he returned as CEO with a new initial board on Nov 21, 2023 and Nov 22, 2023. A recurring case study in AI lab governance.

What it asks to be measured. None.

Source: OpenAI. Checked Oct 08, 2026.

Cite. Protocol STAMP, “OpenAI board removes Sam Altman as CEO; reinstated days later,” The Chronicle, entry ms-openai-board-2023-11-17, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-openai-board-2023-11-17.

Writers Guild of America contract sets rules on AI in screenwriting

The 2023 WGA Minimum Basic Agreement, ending a 148-day strike, provides that neither traditional nor generative AI is a 'writer', so its output is not 'literary material', and unpublished AI-generated material given to a writer is not 'source material' for credit, and requires disclosure when studios hand writers AI-generated material.

What it asks to be measured. Disclosure when AI-generated material is provided to writers.

Source: Writers Guild of America West. Checked Oct 08, 2026.

Limits: The 148-day strike length and ratification date are from contemporaneous reporting, not this page.

Cite. Protocol STAMP, “Writers Guild of America contract sets rules on AI in screenwriting,” The Chronicle, entry us-hollywood-wga-2023-09-27, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#us-hollywood-wga-2023-09-27.

Commission President von der Leyen proposes an IPCC-style global panel for AI (State of the Union 2023)

In her 2023 State of the Union address, the Commission President cited the warning on the risk of extinction from AI and proposed a global scientific panel for AI on the model of the IPCC.

What it asks to be measured. None stated.

“I believe we need a similar body for AI”Ursula von der Leyen, President of the European Commission, official Commission transcript

Source: European Commission (Press corner) · The Peninsula (secondary) (secondary). Checked Oct 08, 2026.

Limits: The Commission page does not print the date; Sep 13, 2023 is from contemporaneous press reports.

See explainer: Landmark speeches on AI.

Cite. Protocol STAMP, “Commission President von der Leyen proposes an IPCC-style global panel for AI (State of the Union 2023),” The Chronicle, entry eu-vdl-soteu-ai-2023-09-13, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#eu-vdl-soteu-ai-2023-09-13.

UN Secretary-General addresses the Security Council's first debate on AI

At the first Security Council debate on artificial intelligence, convened by the United Kingdom, the Secretary-General welcomed calls for a new UN entity to support collective governance of AI, citing bodies such as the International Atomic Energy Agency as models.

What it asks to be measured. None stated.

“a new United Nations entity to support collective efforts to govern this extraordinary technology”António Guterres, UN Secretary-General, official UN transcript

Source: United Nations Secretary-General. Checked Oct 08, 2026.

See explainer: Landmark speeches on AI.

Cite. Protocol STAMP, “UN Secretary-General addresses the Security Council's first debate on AI,” The Chronicle, entry un-guterres-unsc-ai-2023-07-18, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#un-guterres-unsc-ai-2023-07-18.

Mata v. Avianca: lawyers sanctioned for citing fake cases generated by ChatGPT

Judge P. Kevin Castel (S.D.N.Y.) sanctioned two lawyers and their firm USD 5,000 for submitting a brief with non-existent judicial opinions produced by ChatGPT. Led many courts to adopt rules on AI use in filings.

What it asks to be measured. None directly; prompted court disclosure rules for AI-assisted filings.

Source: CourtListener (Free Law Project) docket (secondary). Checked Oct 08, 2026.

Limits: Docket page from a public court-records mirror, not PACER; sanction amount from the opinion as reported.

Cite. Protocol STAMP, “Mata v. Avianca: lawyers sanctioned for citing fake cases generated by ChatGPT,” The Chronicle, entry ct-mata-avianca-2023-06-22, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-mata-avianca-2023-06-22.

Center for AI Safety one-sentence Statement on AI Risk

A one-sentence statement that mitigating the risk of extinction from AI should be a global priority alongside pandemics and nuclear war, signed by leading AI researchers and the heads of major AI labs.

What it asks to be measured. None.

Source: Center for AI Safety. Checked Oct 08, 2026.

Cite. Protocol STAMP, “Center for AI Safety one-sentence Statement on AI Risk,” The Chronicle, entry ms-cais-statement-2023-05-30, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-cais-statement-2023-05-30.

US Senate Judiciary subcommittee hearing 'Oversight of A.I.: Rules for Artificial Intelligence' with Sam Altman

First appearance by OpenAI's chief executive before Congress. Witnesses discussed licensing, independent audits and a possible federal AI agency.

What it asks to be measured. Proposals for licensing and independent audits of powerful models (not enacted).

Source: US Senate Committee on the Judiciary. Checked Oct 08, 2026.

Cite. Protocol STAMP, “US Senate Judiciary subcommittee hearing 'Oversight of A.I.: Rules for Artificial Intelligence' with Sam Altman,” The Chronicle, entry ms-us-senate-hearing-2023-05-16, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-us-senate-hearing-2023-05-16.

Italy's data-protection authority temporarily limits ChatGPT

The Garante ordered OpenAI to stop processing Italian users' data, citing the lack of a legal basis for training-data collection, inaccurate outputs and no age verification. OpenAI suspended service in Italy and restored it in late April 2023 after changes.

What it asks to be measured. None directly; data-protection conditions on model training and age checks.

Source: Garante per la protezione dei dati personali. Checked Oct 08, 2026.

See explainer: Regulators' actions against AI uses.

Cite. Protocol STAMP, “Italy's data-protection authority temporarily limits ChatGPT,” The Chronicle, entry en-it-garante-chatgpt-2023-03-30, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#en-it-garante-chatgpt-2023-03-30.

Future of Life Institute open letter calling for a six-month pause on giant AI experiments

Called on all AI labs to pause for at least six months the training of AI systems more powerful than GPT-4, and to use the pause to develop shared safety protocols audited and overseen by independent outside experts. Signed by tens of thousands; no lab announced a pause.

What it asks to be measured. Proposed independent audits of safety protocols for advanced AI.

Source: Future of Life Institute. Checked Oct 08, 2026.

Cite. Protocol STAMP, “Future of Life Institute open letter calling for a six-month pause on giant AI experiments,” The Chronicle, entry ms-fli-pause-letter-2023-03-22, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-fli-pause-letter-2023-03-22.

Anthropic releases Claude

Anthropic made its Claude assistant available through a chat interface and API after testing with early partners. Released the same day as GPT-4.

What it asks to be measured. None.

Source: Anthropic. Checked Oct 08, 2026.

Cite. Protocol STAMP, “Anthropic releases Claude,” The Chronicle, entry ms-claude-launch-2023-03-14, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-claude-launch-2023-03-14.

US Copyright Office: Zarya of the Dawn registration limited to human-authored elements

The Office cancelled and reissued the registration of Kristina Kashtanova's comic book, covering the text and the selection and arrangement of images but not the individual images generated with Midjourney.

What it asks to be measured. None; human authorship as the test for registration.

Source: US Copyright Office. Checked Oct 08, 2026.

Cite. Protocol STAMP, “US Copyright Office: Zarya of the Dawn registration limited to human-authored elements,” The Chronicle, entry us-fed-copyright-zarya-2023-02-21, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#us-fed-copyright-zarya-2023-02-21.

OpenAI releases ChatGPT

OpenAI released ChatGPT as a 'research preview', a conversational model fine-tuned from the GPT-3.5 series with reinforcement learning from human feedback. Its rapid adoption set off the generative-AI policy wave of 2023.

What it asks to be measured. None; the launch post lists known limitations such as plausible-sounding but incorrect answers.

Source: OpenAI. Checked Oct 08, 2026.

See explainer: Technology milestones that shaped the rules.

Cite. Protocol STAMP, “OpenAI releases ChatGPT,” The Chronicle, entry ms-chatgpt-2022-11-30, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-chatgpt-2022-11-30.

US export controls on advanced computing chips and semiconductor manufacturing items

The Bureau of Industry and Security issued an interim final rule (released Oct 07, 2022, published Oct 13, 2022) restricting exports of advanced AI chips, supercomputer end uses and chip-making equipment to China. Tightened in October 2023 and December 2024.

What it asks to be measured. Performance thresholds for controlled chips; no evaluation duties.

Source: Federal Register. Checked Oct 08, 2026.

Limits: Later tightenings (2023, 2024) are not covered by this source.

Cite. Protocol STAMP, “US export controls on advanced computing chips and semiconductor manufacturing items,” The Chronicle, entry us-fed-bis-chip-controls-2022-10-07, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#us-fed-bis-chip-controls-2022-10-07.

Stable Diffusion released with open weights

Stability AI publicly released the Stable Diffusion text-to-image model. Open release of an image model spurred both adoption and litigation, including Getty Images' UK and US suits.

What it asks to be measured. None.

Source: Stability AI. Checked Oct 08, 2026.

See explainer: Technology milestones that shaped the rules.

Cite. Protocol STAMP, “Stable Diffusion released with open weights,” The Chronicle, entry ms-stable-diffusion-2022-08-22, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-stable-diffusion-2022-08-22.

UK ICO fines Clearview AI GBP 7.5 million

The ICO fined Clearview AI GBP 7,552,800 and ordered it to delete UK residents' data scraped from the internet for its facial-recognition database. The First-tier Tribunal overturned the decision on jurisdiction in 2023; the Upper Tribunal later found for the ICO on that point in 2025.

What it asks to be measured. None directly; limits on scraping biometric data to build recognition systems.

Source: Information Commissioner's Office (Internet Archive copy) (secondary). Checked Oct 08, 2026.

Limits: The ICO notice no longer resolves at either of its ico.org.uk addresses (checked Oct 08, 2026); cited from the Internet Archive capture of Jun 15, 2022. Appeal history from secondary reporting; check the tribunal judgments before relying on it. Sources are copies or reports, not the issuer's own site, so the entry is Likely.

See explainer: Regulators' actions against AI uses.

Cite. Protocol STAMP, “UK ICO fines Clearview AI GBP 7.5 million,” The Chronicle, entry en-uk-ico-clearview-2022-05-23, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#en-uk-ico-clearview-2022-05-23.

GitHub Copilot technical preview

GitHub announced Copilot, an 'AI pair programmer' powered by OpenAI Codex and trained on public source code. It later became the subject of a class action over open-source licences (Doe v. GitHub, filed November 2022).

What it asks to be measured. None.

Source: GitHub. Checked Oct 08, 2026.

Limits: The Doe v. GitHub lawsuit is mentioned from general knowledge, not this source.

See explainer: Technology milestones that shaped the rules.

Cite. Protocol STAMP, “GitHub Copilot technical preview,” The Chronicle, entry ms-github-copilot-2021-06-29, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-github-copilot-2021-06-29.

AlphaFold 2 results at CASP14

DeepMind's AlphaFold 2 achieved a median GDT score of about 92.4 in the CASP14 protein-structure-prediction assessment, a result the organisers described as a solution to a 50-year-old grand challenge for many proteins. The 2021 Nature paper and open database followed.

What it asks to be measured. Blind community assessment (CASP) as an independent benchmark.

Source: Google DeepMind · Nature (Jumper et al. 2021). Checked Oct 08, 2026.

See explainer: Technology milestones that shaped the rules.

Cite. Protocol STAMP, “AlphaFold 2 results at CASP14,” The Chronicle, entry ms-alphafold-casp14-2020-11-30, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-alphafold-casp14-2020-11-30.

Portland bans face recognition by city bureaus and by private entities in public places

Portland City Council approved two ordinances: one banning city bureaus from using face recognition, the other banning private entities from using it in places of public accommodation.

What it asks to be measured. None.

Source: City of Portland. Checked Oct 08, 2026.

Cite. Protocol STAMP, “Portland bans face recognition by city bureaus and by private entities in public places,” The Chronicle, entry us-or-portland-facial-ban-2020-09-09, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#us-or-portland-facial-ban-2020-09-09.

England's exam regulator abandons algorithm-standardised A-level grades

After a statistical model downgraded many teachers' assessed grades, Ofqual announced that students would receive their centre assessment grades instead. A defining public controversy over algorithmic decision-making in the UK.

What it asks to be measured. None.

Source: Ofqual (GOV.UK). Checked Oct 08, 2026.

Cite. Protocol STAMP, “England's exam regulator abandons algorithm-standardised A-level grades,” The Chronicle, entry uk-ofqual-algorithm-2020-08-17, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#uk-ofqual-algorithm-2020-08-17.

IBM exits general-purpose facial recognition; Amazon pauses police use of Rekognition

IBM's CEO wrote to Congress on Jun 08, 2020 that IBM no longer offers general-purpose facial recognition or analysis software and called for testing for bias. On Jun 10, 2020 Amazon announced a one-year moratorium on police use of Rekognition, later extended indefinitely.

What it asks to be measured. Calls for vendor testing of AI systems for bias, audited and reported.

Source: IBM · Amazon. Checked Oct 08, 2026.

Limits: IBM's letter URL now redirects to ibm.com/policy and the text was not retrieved (Internet Archive offline on Oct 08, 2026); IBM details are from contemporaneous reporting. The indefinite extension of Amazon's moratorium is also from reporting.

Cite. Protocol STAMP, “IBM exits general-purpose facial recognition; Amazon pauses police use of Rekognition,” The Chronicle, entry ms-ibm-amazon-facial-recognition-2020-06-08, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-ibm-amazon-facial-recognition-2020-06-08.

GPT-3 paper: 'Language Models are Few-Shot Learners'

OpenAI researchers described GPT-3, a 175-billion-parameter language model, and reported few-shot performance across many benchmarks. OpenAI opened an API to it in June 2020.

What it asks to be measured. Benchmark-based evaluation of few-shot learning; the paper's own discussion of data contamination between training data and benchmarks.

Source: arXiv (Brown et al., OpenAI). Checked Oct 08, 2026.

See explainer: Technology milestones that shaped the rules.

Cite. Protocol STAMP, “GPT-3 paper: 'Language Models are Few-Shot Learners',” The Chronicle, entry ms-gpt3-2020-05-28, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ms-gpt3-2020-05-28.

The Hague District Court strikes down SyRI, the Dutch welfare-fraud risk algorithm

Held that the legislation for SyRI (System Risk Indication), which linked government data to flag people for welfare-fraud investigation, violated the right to private life in Article 8 of the European Convention on Human Rights, citing its lack of transparency.

What it asks to be measured. Transparency and verifiability of a government risk-scoring model (absent here).

Source: Rechtspraak.nl (ECLI:NL:RBDHA:2020:865). Checked Oct 08, 2026.

Cite. Protocol STAMP, “The Hague District Court strikes down SyRI, the Dutch welfare-fraud risk algorithm,” The Chronicle, entry ct-nl-syri-2020-02-05, last checked Oct 08, 2026, https://protocolstamp.com/chronicle/#ct-nl-syri-2020-02-05.