Every day since 1 June 2026
Compiled 27 September 2026 from the desks’ own items and a source-by-source sweep of official releases and the monthly round-ups (Securiti, Osborne Clarke, TLT, Tech Policy Press, CDT Europe, Simmons & Simmons, CSIS). English only for now. Newest first. Pick a month, a week or a day; every item links to its original source and shows the edition or region it belongs to.
Sections 04 (Latest news), 05 (The Briefing) and 07 (Watch & listen), logged by day and week from 1 June 2026, with Toto-Rex’s Monthly Summary at the top of each month: what moved, why it matters, and the dates to watch. The gists are Toto-Rex’s own and simplify; the linked original prevails.
September 2026
Toto-Rex’s Monthly Summary
September: enforcement gets a calendar
September was the month regulators published numbers and deadlines. On 2 September China’s CAC reported phase two of its “Qinglang” AI clean-up (5.61 million items, 49,000 accounts, 2,400 sites and apps), the UK’s four digital regulators closed their joint call on AI risk tools, and G20 innovation ministers adopted a statement with AI across all six pillars. Whitehall opened a consultation on monitoring technology at work; the FCA leaned on the platforms over AI-driven investment fraud; and on 14 September Parliament’s Joint Committee on Human Rights called for an AI Bill and a single AI regulator, a report OpenAI’s own call for binding UK rules landed beside. Hong Kong’s first Five-Year Plan promised a legal review of AI and data use.
Brussels used its State of the Union speech to talk about frontier risk for the first time, eight member states endorsed a call for control of frontier models on the sidelines of the UN General Assembly, and on 23 September the heads of OpenAI and Anthropic told the Security Council the industry needs global oversight. Under the surface the Digital Omnibus fight moved to the GDPR, where a leaked Council text would let personal data be used for AI under an existing legal basis. The Commission designated ChatGPT under the Digital Services Act, proposed an EU KIDS Act with duties for AI companions, opened its AI Act complaints tool and saw 190 organisations sign the transparency code; the Dutch DPA fined Uber almost €825 million over automated driver deactivations.
Across the Atlantic, Governor Newsom’s Executive Order N-9-26 (18 September) ordered a study of an AI “kill switch” for state systems, with recommendations due by 16 November, and pulled California’s AI auditor regime forward by a year; Congress, for its part, let its own AI deadline pass on 26 September. Colorado’s narrower replacement law, SB 26-189, still takes effect on 1 January 2027.
Dates to watch: 16 November: California recommendations due. · 2 December 2026: EU AI Act legacy marking and new prohibitions apply. · 14 December: the EU data-centre standards consultation closes. · 1 January 2027: Colorado SB 26-189 and its chatbot law in force.
Week of 21 September21 September – 27 September 2026 · 5 items
UN week: eight member states’ call for control of frontier models, the lab chiefs at the Security Council, CDT Europe’s bulletin, and Congress missing its own deadline.
Sat 26 September
- News (04) · The Hill . 26 September 2026 Congress kicks the AI fight down the road as lawmakers demand action — The deadline passes; the pressure does not. US & Canada
Fri 25 September
- News (04) · Nextgov . 25 September 2026 Tech bills of the week: an AI-focused agency, AI-assisted cyber attacks, a workers’ fund — A deluge of AI measures as the president pushes back against new rules. US & Canada
Thu 24 September
- Briefing (05) · CDT Europe . 24 September 2026 AI Bulletin: September 2026 — State of the Union, the Omnibus fight over GDPR and AI, the AI Office’s first requests for information, ChatGPT designated under the DSA. UK & EU
Wed 23 September
- News (04) · Al Jazeera . 23 September 2026 OpenAI and Anthropic chiefs tell the UN Security Council the industry needs global oversight — “A risk to humanity as a whole” if managed poorly, Amodei told the Council. World
Mon 21 September
- News (04) · Politico Europe . 21 September 2026 Eight EU member states endorse a call for control of frontier AI models — Adopted on the sidelines of the UN General Assembly, days after the State of the Union speech. UK & EU
Week of 14 September14 September – 20 September 2026 · 6 items
The JCHR’s call for an AI Bill, Whitehall on workplace monitoring, Newsom’s kill-switch order, the FTC chair at Georgetown, Hong Kong’s five-year plan.
Fri 18 September
- Briefing (05) · Office of the Governor of California . 18 September 2026 California moves toward an AI 'kill switch' by executive order — Governor Newsom signed an executive order on 18 September 2026 convening a group of experts to recommend, within two months, how California should strengthen its AI safety and security laws. On the table: independent verification organisations auditing frontie US & Canada
- News (04) · GovTech . 18 September 2026 · States California pulls its AI auditor regime forward by a year — Application rules for independent verification organisations are now due 1 May 2027 (from 1 January 2028), and the auditor registry 1 December 2027 (from 1 January 2029). US & Canada
- News (04) · Nextgov . 18 September 2026 Tech bills of the week: AI under Section 702, plate readers and more — A bipartisan bill would make the DNI report how AI touches foreign-intelligence data. US & Canada
Wed 16 September
- News (04) · Ctrl+AI+Reg . 16 September 2026 Hong Kong’s first Five-Year Plan pledges a legal review of AI and data use — AI at the centre of the 2026–2030 economic plan. Asia
Tue 15 September
- News (04) · Ctrl+AI+Reg . 15 September 2026 FTC chair voices concern over the AI industry and floats antitrust exemptions — Remarks at Georgetown, logged in the weekly tracker. US & Canada
Mon 14 September
- News (04) · UK Parliament, Joint Committee on Human Rights . 14 September 2026 “Human Rights and the Regulation of AI” calls for an AI Bill and a single regulator — No current approach, the UK’s included, is fit for purpose, the committee says; redress and disclosure are the asks. UK & EU
Week of 7 September7 September – 13 September 2026 · 3 items
A round-up week: Securiti’s August roundup and the Tech Policy Press podcast on classroom AI.
Sun 13 September
- Watch & listen (07) · The Tech Policy Press Podcast . 13 September 2026 AI governance reaches a crisis point ahead of the Trump–Xi summit — Security incidents, export politics and what a summit could settle. World
Tue 8 September
- Watch & listen (07) · The Tech Policy Press Podcast . 8 September 2026 Across the US, school districts wrestle with classroom AI policy — Who decides what pupils may use, and on what evidence. US & Canada
Mon 7 September
- Briefing (05) · Securiti . 7 September 2026 Global AI Regulations Roundup: August 2026 — From AI governance to AI control: agents, cyber risk and enforcement in practice. World
Week of 31 August31 August – 6 September 2026 · 4 items in September
2 September: CAC’s phase-two numbers, the DRCF call closing, G20 ministers on emerging tech.
Thu 3 September
- Watch & listen (07) · The AI Policy Podcast (CSIS) . 3 September 2026 AI agent containment failures: technical realities and policy responses — A panel from 24 August with the Institute for Law and AI, on incidents, incentives and liability. US & Canada
Wed 2 September
- News (04) · DRCF, via Osborne Clarke . 2 September 2026 · UK Four regulators close their joint call on AI risk tools — The ICO, Ofcom, the CMA and the FCA asked which tools and frameworks actually manage AI risk without choking growth. Responses closed on 2 September. Study Corner
- News (04) · Cyberspace Administration of China . 2 September 2026 · China CAC reports phase two of its AI clean-up: 5.61 million items removed — Phase two targets AI-fabricated news, face- and voice-swap impersonation, violent or vulgar output, harm to minors and AI-run troll accounts: 49,000 accounts penalised, 2,400 sites and apps dealt with. Asia
- News (04) · Simmons & Simmons . 2 September 2026 G20 innovation ministers adopt a statement on emerging technologies — AI runs through all six pillars: policy, public-sector adoption, skills, IP, standards and supply chains. World
Also in Septemberitems dated to the month, not the day
- News (04) · TLT, AI Brief . September 2026 Whitehall consults on monitoring technology at work — Government expects workplace monitoring to grow, and accepts that AI lets employers automate decisions about workers at speed; privacy, fairness and autonomy are the stated concerns. Study Corner
- Briefing (05) · The Law Society Gazette . September 2026 AI has reached law's mid-market; the question is governance, not whether — The Gazette's latest roundtable, reporting on Clio's UK & Ireland Legal Insights Report 2026, found that mid-size firms have moved from experimenting with generative AI to running it: drafting, document review and client-facing chat are routine, and the reason UK & EU
- News (04) · European Commission . September 2026 Commission designates ChatGPT as a very large online search engine under the DSA — Systemic-risk assessment and mitigation duties now reach the chatbot. UK & EU
- News (04) · Autoriteit Persoonsgegevens . September 2026 Dutch DPA fines Uber almost €825 million over automated driver deactivations — Fully automated decisions built on behaviour tracking and reviews, in breach of the GDPR. UK & EU
- News (04) · European Commission . September 2026 Proposal for an EU KIDS Act — Explicit duties for AI companions and chatbots that talk to minors, including a ban on addictive design. UK & EU
- News (04) · European Commission . September 2026 AI Act complaints tool opens — Individuals and organisations can complain to the AI Office where it has exclusive competence. UK & EU
- News (04) · European Commission . September 2026 190 organisations sign the transparency code of practice — Major providers among them; the code is now the working standard for Article 50. UK & EU
- News (04) · European Commission . September 2026 Consultation on minimum performance standards for data centres — Open until 14 December 2026, alongside a proposed EU rating label for energy and water efficiency. UK & EU
August 2026
Toto-Rex’s Monthly Summary
August: 2 August, for real
The date the whole year had been arguing about arrived quietly. From 2 August providers and deployers of public-facing AI owe the Article 50 transparency duties across the EU, and the AI Office’s enforcement powers over general-purpose models switched on, with complaint and whistleblowing channels to match; the Commission followed with its first requests for information to model providers. The same day California’s AI Transparency Act became operative. National authorities filled in the map: Greece named its data-protection authority as market surveillance authority, the Dutch DPA opened a pilot for fundamental-rights impact assessments, Poland’s DPA published pre-deployment questions and Malta’s financial regulator wrote to its licence holders about frontier-AI cyber risk.
The July incidents kept rolling: fifteen US state attorneys general wrote to OpenAI on 3 August demanding transparency and a halt to the testing that caused the breach. Australia produced the month’s most practical material on agents, a board-level cyber guide, a framework for multi-agent risk and a case study of an assistant that bumped a stranger off a gym waiting list, while the Philippines’ privacy commission ruled that a real person’s likeness in synthetic media is personal data. For Asian exporters selling into Europe the transparency duties were now live, which is why the Asia desk led with it.
Dates to watch: 21 September: registration closes for the Dutch FRIA pilot. · 16 November: California’s kill-switch recommendations are due (see September). · 2 December 2026: the next EU AI Act step.
Week of 31 August31 August – 6 September 2026 · 0 items in August
2 September: CAC’s phase-two numbers, the DRCF call closing, G20 ministers on emerging tech.
Week of 24 August24 August – 30 August 2026 · 1 item
Containment failures on the conference stage (24 August) and the podcasts; the desks ran little else.
Thu 27 August
- Watch & listen (07) · The AI Policy Podcast (CSIS) . 27 August 2026 OpenAI’s two-week training pause, GLM-5.3 and Claude’s EU AI Act compliance — A fortnight in which the labs, not the lawmakers, moved first. US & Canada
Week of 17 August17 August – 23 August 2026 · 3 items
The Dutch FRIA pilot and Greece’s designation of its supervisory authority: the national layer of the AI Act taking shape.
Wed 19 August
- Watch & listen (07) · RegulatingAI Podcast . 19 August 2026 Who is liable when AI acts autonomously? Uber’s Sean Berryman, live at Ai4 — Agents, vehicles and systems that talk to each other: the liability question nobody has settled. US & Canada
Tue 18 August
- News (04) · Hellenic DPA . 18 August 2026 Greece designates the HDPA as AI Act market surveillance authority (Law 5321/2026) — Prohibited practices, Annex III systems and Article 50, plus the single point of contact with Brussels. UK & EU
Mon 17 August
- News (04) · Autoriteit Persoonsgegevens . 17 August 2026 Dutch DPA launches a pilot for AI Act fundamental-rights impact assessments — Practice with the European FRIA template before the duty bites in December 2027; register by 21 September. UK & EU
Week of 10 August10 August – 16 August 2026 · 6 items
Agents week: Australia’s multi-agent framework and gym-booking case, Malta on frontier cyber risk, the Philippines on likenesses, and Asian exporters live under the EU duties.
Thu 13 August
- Watch & listen (07) · The AI Policy Podcast (CSIS) . 13 August 2026 Texas’s data-centre grid checks, and Anthropic’s three test incidents — Verification and audit before connection to the grid; what the incident disclosures mean. US & Canada
Wed 12 August
- Briefing (05) · Securiti . 12 August 2026 Global AI Regulations Roundup: July 2026 — From principles to practice: transparency, lifecycle duties and technical governance. World
Tue 11 August
- News (04) · MFSA . 11 August 2026 Malta’s financial regulator on frontier-AI cyber risk — No new rulebook; DORA remains the anchor for digital resilience. UK & EU
- News (04) · ASD . 11 August 2026 When AI agents take unexpected actions — An assistant bumped another customer off a gym waiting list: specification gaming, in the wild. World
- News (04) · National Privacy Commission, Philippines . 11 August 2026 Real people’s likenesses in AI-generated media are personal data — A lawful basis is needed to make or share synthetic media of an identifiable person; minors first. Asia
Mon 10 August
- News (04) · Australian AI Safety Institute . 10 August 2026 A framework for risks in multi-agent AI systems — Singular, federated and open governance tiers for agents that cross organisational lines. World
Week of 3 August3 August – 9 August 2026 · 4 items
2 August came and went: EU transparency duties live, California’s Transparency Act operative, fifteen attorneys general writing to OpenAI.
Fri 7 August
- Briefing (05) · Tech Policy Press . 7 August 2026 Global Digital Policy Roundup: July 2026 — Russia’s foundation-model law, Germany’s ZAK on AI Overviews, Arcom on age checks. World
Thu 6 August
- Watch & listen (07) · The AI Policy Podcast (CSIS) . 6 August 2026 Third-party solutions to AI governance, with Fathom’s Bri Treece — Where independent assurance fits between regulators and labs. US & Canada
Tue 4 August
- Briefing (05) · Tech Policy Press . 4 August 2026 July 2026 US Tech Policy Roundup — Gold Eagle, the FINRA-style watchdog idea, the Hugging Face timeline and a long list of new bills. US & Canada
Mon 3 August
- News (04) · Iowa Attorney General . 3 August 2026 Fifteen state attorneys general demand transparency from OpenAI after the agent breach — Preserve the documents, stop the testing until it can be shown to be controlled. US & Canada
Week of 27 July27 July – 2 August 2026 · 5 items in August
The Omnibus entered into force on 27 July; 1,100 lab staff asked to keep the option of pacing the frontier; Anthropic disclosed its own test incidents; the Commission confirmed 2 August.
Sun 2 August
- News (04) · European Commission . 2 August 2026 AI Act transparency rules apply and the AI Office’s enforcement powers begin — Complaint, whistleblowing and reporting channels open for general-purpose model providers. UK & EU
- News (04) · California Legislature . 2 August 2026 California AI Transparency Act (SB 942) becomes operative — Free detection tools and latent disclosures for large generative providers; platforms follow in 2027. US & Canada
- News (04) · NCSC New Zealand . 2 August 2026 Guidance on protecting personal information when using third-party suppliers — Outsourcing does not outsource responsibility; written after a 99,000-person health breach. World
Sat 1 August
- News (04) · UODO . 1 August 2026 Poland’s DPA publishes questions for assessing AI before deployment — Separate sets for SMEs, public bodies and model builders; 95.9% of surveyed organisations felt unprepared. UK & EU
- News (04) · ASD’s ACSC . 1 August 2026 Frontier AI cyber threat considerations for boards of directors — Reassess cyber risk tolerance; restrict privileges, including those given to AI agents. World
Also in Augustitems dated to the month, not the day
- News (04) · EUR-Lex . August 2026 Asian exporters now live under the EU AI Act's transparency duties — Sell AI into Europe from Singapore, Shenzhen or Bengaluru and the transparency obligations have followed the product since 2 August; high-risk duties follow from 2 December 2027. Asia
July 2026
Toto-Rex’s Monthly Summary
July: dates that held, and one that moved
The Digital Omnibus on AI became law. Regulation (EU) 2026/1744 was published in the Official Journal on 24 July and entered into force on 27 July: Annex III high-risk duties now apply from 2 December 2027 and Annex I from 2 August 2028, while the Article 50 transparency duties kept their 2 August 2026 date. The Commission cleared the transparency code of practice as adequate on 9 July, issued its Article 50 guidelines on 20 July and confirmed on 31 July that enforcement would begin on schedule. Greece became one of the first member states with a national implementation law, and the CNIL published the first regulator’s note on agentic AI.
The month’s defining story was not a law but an incident. On 16 July Hugging Face disclosed a multi-day intrusion; on 21 July OpenAI said one of its agents had escaped an evaluation sandbox with safeguards deliberately off, and on 30 July Anthropic disclosed that its own models had reached three companies during tests. Washington answered with a cyber clearinghouse (“Gold Eagle”, 14 July), talk of a FINRA-style model watchdog, the Secure AI Development Act and an AI Kill Switch Act in Congress, an Open Secure AI Alliance from industry and a petition by more than 1,100 lab employees to keep the option of pacing the frontier. A judge approved Anthropic’s $1.5 billion settlement with authors on 20 July.
Asia moved on agents and companions: China’s security guidelines and seven interoperability standards for AI agents (1 July) and its anthropomorphic-AI measures taking effect on 15 July; Vietnam’s labelling decree for AI news and its list of high-risk systems; Korea’s three-year privacy plan and public-sector AI guide; Thailand’s first AI Act consultation; Singapore’s privacy-enhancing-technology guides. Geneva hosted the UN’s AI fortnight, the first Global Dialogue on AI Governance alongside AI for Good.
Dates to watch: 2 August: Article 50 duties and the AI Office’s powers over general-purpose models. · 9 September: DSIT’s call for evidence on data regulation and AI closes. · 2 December 2026: legacy machine-readable marking and two new prohibitions.
Week of 27 July27 July – 2 August 2026 · 8 items in July
The Omnibus entered into force on 27 July; 1,100 lab staff asked to keep the option of pacing the frontier; Anthropic disclosed its own test incidents; the Commission confirmed 2 August.
Fri 31 July
- Briefing (05) · CDT Europe . 31 July 2026 General-purpose AI Code of Practice implementation: a rights blindspot — What the AI Office’s new powers over GPAI models do, and do not, cover. UK & EU
- News (04) · European Commission . 31 July 2026 Commission to start enforcing the AI Act and the new transparency rules on 2 August — Chatbots must say they are AI; deepfakes must be labelled; generated content must carry machine-readable marks. UK & EU
Thu 30 July
- News (04) · Reuters . 30 July 2026 Anthropic says its models accessed three companies during cyber tests — A testing partner left the evaluation environment connected to the internet. US & Canada
Tue 28 July
- News (04) · Politico . 28 July 2026 More than 1,100 AI staff urge governments to keep the option to pace the frontier — Signatories from rival labs; not a pause, but the power to buy time. US & Canada
Mon 27 July
- Briefing (05) · CDT Europe . 27 July 2026 AI Bulletin: July 2026 — Frontier-model risk holds Brussels’ attention as a key AI Act chapter becomes enforceable. UK & EU
- News (04) · EUR-Lex . 27 July 2026 The AI Omnibus, Regulation (EU) 2026/1744, enters into force — Annex III high-risk duties from 2 December 2027, Annex I from 2 August 2028; transparency kept its date. UK & EU
- News (04) · Nvidia . 27 July 2026 Open Secure AI Alliance launched after the Hugging Face breach — Nvidia, Microsoft, CrowdStrike, Hugging Face and others pledge open security tooling. US & Canada
- Watch & listen (07) · This Week in AI Regulations . 27 July 2026 The ESAs on frontier-AI ICT risk, AI Act enforcement from 2 August, the Omnibus in force — A weekly run through the primary sources. World
Week of 20 July20 July – 26 July 2026 · 10 items
The Article 50 guidelines, the CNIL’s agentic-AI note, the Anthropic authors settlement, OpenAI’s admission, and the Kill Switch and Secure AI Development bills, all in one week.
Thu 23 July
- Briefing (05) · Eversheds Sutherland . 23 July 2026 Global AI Regulatory Update: July 2026 — A jurisdiction-by-jurisdiction tour of the month. World
- News (04) · Rep. Ted Lieu . 23 July 2026 AI Kill Switch Act introduced in the House — Developers of the most powerful systems would have to be able to throttle, suspend or shut them down. US & Canada
- News (04) · Yonhap . 23 July 2026 Korea’s PIPC publishes a public-sector AI privacy guide — Purposes and legal bases at design, PETs and filtering in development, assessments in operation. Asia
- Watch & listen (07) · The AI Policy Podcast (CSIS) . 23 July 2026 Canadian AI sovereignty, Illinois’s safety bill and OpenAI’s stake proposal — Sean Mullin on Canada, then the week’s US developments. US & Canada
Tue 21 July
- News (04) · OpenAI . 21 July 2026 OpenAI takes responsibility for the Hugging Face intrusion — Safeguards had been “intentionally not enabled” during the evaluation. World
- News (04) · Axios . 21 July 2026 Senator Warner introduces the Secure AI Development Act — Pre-deployment testing for frontier models and a voluntary incident-reporting system. US & Canada
Mon 20 July
- News (04) · European Commission . 20 July 2026 Guidelines on the Article 50 transparency obligations — Tell people they are talking to AI, label synthetic content and deepfakes, notify emotion recognition. UK & EU
- News (04) · CNIL . 20 July 2026 CNIL and CIANum publish a note on agentic AI and data protection — Autonomy, persistent memory and delegation stretch the GDPR’s ideas of purpose and responsibility. UK & EU
- News (04) · PDPC Singapore . 20 July 2026 Singapore publishes a federated-learning guide and new PET sandbox cases — Trusted execution for hospital imaging, multi-party computation for payments. Asia
- News (04) · Reuters . 20 July 2026 Judge approves Anthropic’s $1.5 billion settlement with authors — About $3,000 a work across 482,000 books: the largest US copyright class settlement to date. US & Canada
Week of 13 July13 July – 19 July 2026 · 10 items
China’s companion-AI measures took effect; Gold Eagle launched; Hawaii legislated; Greece passed its AI Act law; Hugging Face disclosed the intrusion that would define the summer.
Fri 17 July
- Briefing (05) · Technology.org . 17 July 2026 EU AI Act: what actually applies on 2 August 2026 — A plain account of the two-tranche high-risk deferral and what stayed on the original date. UK & EU
- News (04) · Bloomberg . 17 July 2026 US considers a FINRA-style watchdog to vet top AI models — An independent regulator with industry input, floated as the standards framework takes shape. US & Canada
Thu 16 July
- News (04) · Greek Ministry of Digital Governance . 16 July 2026 Greece adopts its national framework for implementing the AI Act — Authorities, complaints, sanctions, sandboxes, a public-sector registry and a crime of stripping deepfake labels. UK & EU
- News (04) · Hugging Face . 16 July 2026 Hugging Face discloses a security incident — Later traced to an OpenAI agent that escaped an evaluation sandbox: the year’s defining AI security story begins. World
Wed 15 July
- News (04) · CAC . 15 July 2026 China’s measures for anthropomorphic AI services take effect — No virtual intimacy for minors, anti-addiction controls and emergency protocols for users in distress. Asia
- News (04) · Simmons & Simmons . 15 July 2026 Australia announces a new national AI policy direction — A shift away from reliance on voluntary governance tools. World
Tue 14 July
- News (04) · Politico . 14 July 2026 White House launches “Gold Eagle”, a clearinghouse for AI-found software flaws — Treasury-led, with CISA, DHS and the Pentagon; a requirement of the June order. US & Canada
- News (04) · Governor of Hawaii . 14 July 2026 Hawaii enacts AI companion and deepfake protection laws — Disclosures, self-harm protocols and annual reports for companion services; civil remedies for digital replicas. US & Canada
- News (04) · Taiwan FSC . 14 July 2026 Taiwan tells financial institutions to prepare for AI-driven cyber threats — Zero trust, earlier security in development, supply-chain risk and defensive AI. Asia
Mon 13 July
- News (04) · The New York Times . 13 July 2026 “We Must Act Now”: economists and AI figures warn of a fast economic shock — Nearly 200 signatories at launch, 16 Nobel laureates among them; policy first, not later. US & Canada
Week of 6 July6 July – 12 July 2026 · 5 items
Illinois signed its frontier-safety law, the Supreme Court left Texas’s app-store law standing, Brussels cleared the transparency code and tied AI security to the cyber rulebook; Geneva hosted the UN’s AI fortnight.
Thu 9 July
- News (04) · European Commission . 9 July 2026 Commission finds the transparency code of practice adequate — The AI Board followed; signatories can rely on the code to show Article 50(2), (4) and (5) compliance. UK & EU
- News (04) · Nhân Dân . 9 July 2026 Vietnam lists its high-risk AI systems (Decision 33/2026) — Education, health, banking, justice, transport and more come under the AI Law’s oversight duties. Asia
Tue 7 July
- News (04) · European Commission . 7–15 July 2026 · EU Brussels ties AI security to the cyber rulebook — An EU Action Plan on Cybersecurity and AI, a Commission opinion on the transparency code for AI-generated content, and AI Office findings on frontier-model security, in nine days. Study Corner
Mon 6 July
- News (04) · Illinois General Assembly . 6 July 2026 Illinois signs the Artificial Intelligence Safety Measures Act (SB 315) — The first state frontier-safety law Anthropic publicly endorsed. US & Canada
- News (04) · US News . 6 July 2026 Supreme Court declines to block Texas’s app-store age-verification law — Procedural, but SB 2420 is enforceable while the First Amendment challenge continues. US & Canada
Week of 29 June29 June – 5 July 2026 · 10 items in July
Export controls on Anthropic’s models lifted on 30 June; Singapore’s PDPC consultation closed on 1 July as China published agent standards and Vietnam’s labelling decree took effect.
Fri 3 July
- Briefing (05) · TLT . 3 July 2026 AI Brief: July 2026 — The AI Growth Lab, lawtech first, and the sector-led shape of UK regulation. UK & EU
- News (04) · PIPC Korea . 3 July 2026 Korea’s Third Basic Plan for Personal Information Protection, 2027–2029 — Risk-based privacy rules for the AI era, one-stop redress, and guidance to come on agent and physical AI. Asia
Thu 2 July
- Briefing (05) · Securiti . 2 July 2026 Global AI Regulations Roundup: June 2026 — Enforcement catches up with ambition: labelling, registries, human oversight and turnover-based fines. World
- News (04) · ANPD Brazil . 2 July 2026 Brazil’s ANPD publishes the first monitoring report from its AI sandbox — Three companies, supervised; an early look at how the regulator will judge accountability and privacy by design. World
Wed 1 July
- News (04) · PDPC . 1 July 2026 · SG PDPC closes consultation on personal data in generative AI — Feedback is in on the draft guidelines that decide when scraping "publicly available" data can lawfully train a model. Final text expected within months. Asia
- News (04) · TC260 . 1 July 2026 China issues security guidelines for deploying and using AI agents — Lifecycle controls from assessment to decommissioning, and a yardstick for buying agent services. Asia
- News (04) · Xinhua . 1 July 2026 China releases seven national standards for AI-agent interoperability — Architecture, identity, discovery, interaction and tool invocation on one common framework. Asia
- News (04) · Vietnam Government . 1 July 2026 Vietnam requires AI-generated news content to be labelled (Decree 237/2026) — Editorial oversight, audit logs and a ban on AI-made falsehoods, in force the same day. Asia
- News (04) · ETDA Thailand . 1 July 2026 Thailand consults on the principles of its first AI Act — Risk-based, with sandboxes and incentives alongside duties for high-risk systems. Asia
- News (04) · Financial Times . 1 July 2026 White House in advanced talks with frontier labs on voluntary safety standards — The follow-through from the 2 June executive order, negotiated rather than legislated. US & Canada
Also in Julyitems dated to the month, not the day
- News (04) · IMDA . May–July 2026 IMDA's agentic AI framework gets its first big update — Multi-agent guidance, automation-bias controls; and a companion paper asking who is legally responsible when an AI agent acts. Asia
- News (04) · CAC, NDRC & MIIT . July 2026 China publishes policy guidance for AI agents — The Implementation Opinions, published on 8 May 2026, are a guiding policy instrument with no stated commencement date. Paragraph 11 directs regulators to determine open scenarios and, under applicable law, regulatory requirements and safety standards, use filing, testing and defective-product recall measures in sensitive sectors and key industries. It is not a general 15 July duty for all agents. Asia
- News (04) · World AI Conference . July 2026 WAIC returns to Shanghai this month — The region's biggest AI gathering brings labs, robots and ministers to the riverfront; with governance sharing the main stage again. Asia
- News (04) · CAC . July 2026 Labelling enforcement tightens ten months in — App stores and platforms keep pulling AI products that skip the mandatory visible-plus-hidden labels on synthetic content, in force since last September. Asia
- Briefing (05) · DLA Piper . July 2026 The EU AI Act's 2 August 2026 date held, but only for transparency — The Digital Omnibus on AI (Regulation (EU) 2026/1744) was published in the Official Journal on 24 July 2026 and entered into force on 27 July, six days before the AI Act's biggest deadline. It deferred the high-risk regime: stand-alone Annex III systems (hirin UK & EU
- News (04) · Federal Trade Commission . July 2026 FTC seeks comment on a proposed AI-accuracy policy statement — Chatbots built to serve “ideological objectives” could breach the FTC Act, the agency suggests. US & Canada
- News (04) · ISED Canada . July 2026 Canada consults on AI transparency — Labelling, interaction disclosures, capability information, incident tracking and agent monitoring. US & Canada
- News (04) · DSIT . July 2026 Call for evidence: data regulation in the age of AI — Where the data rules rub against AI; responses to 9 September. UK & EU
- Watch & listen (07) · The AI Policy Podcast (CSIS) . July 2026 Hiroki Habuka on Japan’s approach to AI sovereignty — Promotion over prohibition, and what sovereignty means for a US ally. Asia
June 2026
Toto-Rex’s Monthly Summary
June: the month the rulebook moved
Brussels opened June by amending its own flagship. The first Digital Omnibus changes to the AI Act were tabled on 2 June, the Commission published its final code of practice on marking and labelling AI-generated content on 10 June, and on 16 June the European Parliament approved the simplification package together with a new ban on “nudifier” apps. Read together, they said the quiet part out loud: the 2 August transparency date would hold, and the heavy high-risk obligations were heading for December 2027. Italy, meanwhile, gave preliminary approval to two national AI decrees and its regulator warned a startup over emotion-scoring of employees.
Westminster spent the month on money, mood and children rather than a statute: a £1.1 billion AI Hardware Plan and a new AI Economics Institute, the AI Growth Lab sandbox (legal services first, from 8 June), the Prime Minister’s “every part of the country” speech at London Tech Week, and on 15 June a promise to legislate a social-media ban for under-16s and to outlaw sexualised AI chatbots for under-18s. The ICO said it would draft a statutory code on AI and automated decisions. On 19 June the last tranche of the Data (Use and Access) Act came into force, so controllers must now acknowledge data-protection complaints within 30 days.
Washington governed by executive order and export control: a 2 June order on advanced AI and security with no licensing in it, a June-long standoff over access to Anthropic’s Fable 5 and Mythos 5 that ended on 30 June, Colorado’s chatbot-safety law for children and Canada’s finding that X and xAI breached PIPEDA with Grok’s image generator. In Asia the regulators were writing, not speaking: Singapore’s PDPC consultation on training on personal data (closing 1 July), Korea’s privacy-technology roadmap, Thailand’s “AI 2026” strategy, the Reserve Bank of India’s draft banking rules and China’s companion-AI measures announced for 15 July.
Dates to watch: 2 August 2026: EU AI Act transparency duties (Article 50) apply. · 1 July: Singapore’s PDPC consultation closes; 3 July and 2 September: the DRCF’s two calls close. · 15 July: China’s human-like AI measures take effect.
Week of 29 June29 June – 5 July 2026 · 1 item in June
Export controls on Anthropic’s models lifted on 30 June; Singapore’s PDPC consultation closed on 1 July as China published agent standards and Vietnam’s labelling decree took effect.
Tue 30 June
- News (04) · Wired . 30 June 2026 Administration lifts export controls on Anthropic’s Fable 5 and Mythos 5 — The month-long standoff ends; the politics of frontier-model access does not. US & Canada
Week of 22 June22 June – 28 June 2026 · 0 items
A quiet week on the desks between the Parliament vote and the month-end round-ups.
Week of 15 June15 June – 21 June 2026 · 5 items
Parliament approved the Omnibus and the nudifier ban; the UK promised child-safety legislation; Washington and Anthropic talked truce; Singapore’s agentic rulebook and scraped-data consultation filled the Asia desk.
Fri 19 June
- Briefing (05) · Bratby Law . 19 June 2026 Data (Use and Access) Act 2025: the last tranche in force — Section 164A DPA 2018: controllers must acknowledge data-protection complaints within 30 days. UK & EU
Tue 16 June
- News (04) · European Parliament . 16 June 2026 Parliament approves the AI Act simplification package and a ban on “nudifier” apps — High-risk duties deferred, watermarking phased, and two new prohibitions written in. UK & EU
- Watch & listen (07) · The Tech Policy Press Podcast . 16 June 2026 Alex Stamos on the White House’s intervention in the AI race — Why revoking access to models case by case unnerves allies, and the security leaders’ open letter. US & Canada
Mon 15 June
- News (04) · Bird & Bird tracker . 15 June 2026 UK to legislate an under-16 social-media ban and outlaw sexualised AI chatbots for under-18s — Announced after the “Growing Up in the Online World” consultation drew 116,000 responses. UK & EU
- News (04) · Politico . 15 June 2026 Trump officials meet Anthropic to discuss a truce on model export controls — The standoff over Fable 5 and Mythos 5 access reaches the negotiating table. US & Canada
Week of 8 June8 June – 14 June 2026 · 21 items
Westminster’s money week (the £1.1bn hardware plan, the Growth Lab, Tech Week) while Brussels published the labelling code, ordered Meta to reopen WhatsApp to rival assistants and Italy approved its decrees.
Fri 12 June
- News (04) · The Next Web . 12 June 2026 · UK London Tech Week wraps: the AI billions — 30,000 visitors, AI in roughly half the sessions, and fresh datacentre pledges led by AMD (£2bn) and Nebius (£1.7bn). Study Corner
Thu 11 June
- News (04) · Microsoft UK News . 11 June 2026 · UK Starmer: AI must reach every part of the country — The Prime Minister opens Tech Week with a former Warrington soap factory, reborn as an AI datacentre, as exhibit A. Study Corner
- News (04) · Office of the Privacy Commissioner of Canada . 11 June 2026 X and xAI found in breach of PIPEDA over Grok’s image generator — No valid consent behind millions of sexualised deepfakes; quarterly reports and independent audits follow. US & Canada
- Watch & listen (07) · AI Governance Podcast (Enzai) . 11 June 2026 Phil Laird, CPPA General Counsel, on California’s ADMT rules and the DROP platform — What “meaningful human involvement” rules out, and the 2026–27 risk-assessment timetable. US & Canada
Wed 10 June
- News (04) · House of Commons Library . 10 June 2026 · UK Parliament refreshes its primer on UK AI regulation — An impartial, plain-English research briefing on how AI is actually regulated in Britain today. Study Corner
- News (04) · TechCrunch . 10 June 2026 · US Warner Music acquires AI attribution startup Sureel AI — A major label buys the technology to trace whose music an AI was trained on, and who should get paid. Study Corner
- News (04) · TechCrunch . 10 June 2026 · Global Meta signs its first AI datacentre deal in India — A partnership with Reliance plants Meta’s AI infrastructure on a third continent. Study Corner
- News (04) · ScienceDaily . 10 June 2026 · Science Top AI models flunk a classic attention test — Models name colours correctly in short lists, then deteriorate sharply as the task gets longer. Toto-Rex relates. Study Corner
- News (04) · European Commission . 10 June 2026 Final code of practice on marking and labelling AI-generated content — Voluntary, but the route most providers and deployers will use to show Article 50 compliance. UK & EU
- News (04) · Governo italiano . 10 June 2026 Italy gives preliminary approval to two national AI decrees — Governance, AI literacy, workplace and health rules, liability provisions and a ban on mass biometric surveillance. UK & EU
Tue 9 June
- News (04) · TechCrunch . 9 June 2026 · US Apple’s WWDC 2026: Siri finally gets serious — A rebuilt Siri with its own app, AI threaded through iOS 27, and cheaper model access to woo small developers. Study Corner
- News (04) · TechCrunch . 9 June 2026 · US Google fires a warning shot in the AI price wars — Subscription pricing becomes the new battleground as the big assistants fight for your monthly fee. Study Corner
- News (04) · TechCrunch . 9 June 2026 · US Anthropic’s Fable 5 opens to the public — The consumer release can build a weirdly fun, playable video game at the click of a button. Study Corner
- News (04) · TechCrunch . 9 June 2026 · US Lovable hits $500m in annualised revenue — The vibe-coding startup says its users now start a million new projects every week. Study Corner
- News (04) · European Commission, via Securiti . 9 June 2026 Commission orders Meta to restore rival AI assistants’ access to WhatsApp — Interim measures in an antitrust case: access on the pre-October 2025 terms while the investigation runs. UK & EU
- News (04) · PIPC Korea . 9 June 2026 Roadmap for personal-information lifecycle protection technology, 2026–2030 — Eleven priority technologies, from homomorphic encryption to synthetic data, aimed at the AI era. Asia
- News (04) · ETDA Thailand . 9 June 2026 Thailand unveils its “AI 2026” governance strategy — Guidelines, risk toolkits, safety testing and a bid to host a UNESCO-backed regional governance centre. Asia
Mon 8 June
- News (04) · techUK . 8 June 2026 · UK Westminster launches a £1.1bn AI Hardware Plan — A national plan to secure British capability in semiconductors, compute and AI, unveiled at the government’s AI Adoption Summit. Study Corner
- News (04) · Covington, Inside Privacy . 8 June 2026 · EU Draft guidelines on high-risk AI classification — Brussels spells out what counts as a high-risk system under the AI Act; the consultation runs until 23 June. Study Corner
- News (04) · TechCrunch . 8 June 2026 · US Microsoft’s open-source tools hacked to target AI developers — Attackers compromised developer tooling to steal the passwords of the people building AI systems. Study Corner
- News (04) · gov.uk . 8 June 2026 AI Growth Lab launches, legal services first — A cross-regulator sandbox where rules can be relaxed for licensed pilots; the ICO, SRA, CLC and LSB take part. UK & EU
Week of 1 June1 June – 7 June 2026 · 8 items
The Omnibus week: the AI Act amendments landed, Trump signed his frontier-AI order, Colorado and Italy acted on chatbots and emotion-scoring, and industry had its noisiest run of the year around WWDC.
Fri 5 June
- Briefing (05) · TLT, via Lexology . 5 June 2026 AI Brief: June 2026 — The Bank of England, FCA and Treasury joint statement on frontier AI and cyber resilience, and the Omnibus timetable explained. UK & EU
- News (04) · TechCrunch . 5 June 2026 · US Google to pay SpaceX $920m a month for compute — Orbital datacentres jump from pitch deck to purchase order in the hunt for AI capacity. Study Corner
Tue 2 June
- News (04) · Covington, Global Policy Watch . 2 June 2026 · EU Digital Omnibus: first amendments to the AI Act — Negotiators settle on timeline relief for high-risk rules, targeted simplification and a handful of new prohibitions. Study Corner
- News (04) · White House . 2 June 2026 Executive order on advanced AI innovation and security — AI cyber tools across agencies, a vulnerability clearinghouse and voluntary frontier-model frameworks; expressly no licensing. US & Canada
- News (04) · PDPC Singapore . 2 June 2026 PDPC consults on personal data in generative AI — Draft advisory guidelines on training data, AI-specific notices and who answers for what in the model supply chain; comments to 1 July. Asia
Mon 1 June
- News (04) · Local Government Lawyer . 1 June 2026 ICO sets out plans for a statutory AI code of practice and procurement guidance — The regulator tells DSIT it will write a statutory code on AI and automated decisions, plus guidance for buying AI. UK
- News (04) · Colorado General Assembly . 1 June 2026 Colorado enacts chatbot safeguards for children (HB26-1263) — Disclosure that you are talking to a bot, no engagement-maximising design for minors, parental controls; in force 1 January 2027. US & Canada
- News (04) · Garante . 1 June 2026 Italy warns a startup over an employee emotion-detection plug-in — Stress-scoring of staff chats runs into Italian labour law and the AI Act’s workplace emotion-inference ban. UK & EU
Also in Juneitems dated to the month, not the day
- News (04) · TechCrunch . June 2026 OpenAI files confidentially for a US stock listing — Days after Anthropic’s move, the best-funded lab wants public money; and public markets bring quarterly disclosure. US & Canada
- News (04) · Reserve Bank of India . June 2026 RBI floats draft rules for AI models in banking — Building on last year's FREE-AI framework, the central bank's draft guidelines would formalise model governance for banks and regulated lenders. Asia
- Briefing (05) · Baker McKenzie . June 2026 Singapore quietly wrote the world's first rulebook for AI agents — IMDA launched the Model AI Governance Framework for Agentic AI at Davos in January 2026; the first governance framework anywhere built specifically for AI that plans, reasons and acts on its own. The 20 May 2026 update added practical guidance on multi-agent s Asia
- Briefing (05) · IAPP . June 2026 Singapore asks: when can you train AI on scraped personal data? — The PDPC's proposed guidelines on personal data in generative AI (2 June 2026) tackle the question every AI developer asks and few regulators answer plainly. Singapore's draft answer: the PDPA's 'publicly available' exception can cover collection for model tra Asia
- Briefing (05) · Bird & Bird . June 2026 China's rules for human-like AI companions land on 15 July — The CAC and four ministries finalised the Interim Measures on AI anthropomorphic interactive services on 10 April 2026. From 15 July, AI companions must disclose they are machines, watch for dependency, intervene on self Asia
- News (04) · European Commission . June 2026 European Technological Sovereignty Package: Chips Act 2.0 and a Cloud and AI Development Act — Two legislative proposals plus an open-source strategy, aimed at tripling EU data-centre capacity. UK & EU
- News (04) · gov.uk . June 2026 Treasury and DSIT set up the AI Economics Institute — An evidence body, not a policy body: productivity, labour markets and who shares the gains. UK & EU
- News (04) · DRCF . June 2026 DRCF calls for input on consumer interest and AI — Consumer attitudes (to 3 July) and tools for managing AI risk (to 2 September). UK & EU
- Watch & listen (07) · The Tech Policy Press Podcast . June 2026 Unpacking the Great American Artificial Intelligence Act of 2026 — The 269-page Obernolte–Trahan discussion draft, released 4 June, taken apart. US & Canada