# Policy [https://timesof.ai/category/policy](https://timesof.ai/category/policy) Regulation, export controls, government programs, and public-sector AI decisions. ## Coverage - [UK AI Minister Says an AI-Driven Breach of Britain's Defences Is 'Likely, if Not Inevitable'](https://timesof.ai/2026/10/uk-narayan-ai-cyber-breach-likely-frontier-ai-defences) — UK AI Minister Kanishka Narayan says an AI-driven breach is 'likely, if not inevitable,' orders frontier AI onto government systems; law changes only assessed. - [White House Super Intelligence Force Says AI Incident Reporting Is 'Not Optional,' but Publishes No Rule](https://timesof.ai/2026/10/white-house-si-force-ai-incident-reporting-not-optional) — The SI Force told Axios AI firms must report and fix incidents and that this is 'not optional.' No order, rule, incident definition or penalty has been issued. - [SEBI Says AI Guidelines for India's Capital Markets Are Coming, Built Around Accountability, Kill Switches and Human Oversight](https://timesof.ai/2026/10/sebi-ai-ml-guidelines-capital-markets-kill-switches) — SEBI's Pandey says AI/ML guidelines for capital markets are coming soon: tiered, with kill switches and human oversight. No date; SEBI consulted on a draft in 2025. - [Nadella: Treat AI Models as Insider Risks, With an Emergency Brake and Seven Containment Principles](https://timesof.ai/2026/10/nadella-models-as-insider-risks-seven-principles) — Satya Nadella's Oct. 10 essay says AI models should be contained like insider risks, with an emergency brake, tamper-resistant logs and incident disclosure. - [Unredacted Kentucky Lawsuit Sets Out Chatbot Conversations the State Says Show Character.AI Harming Users](https://timesof.ai/2026/10/kentucky-ag-character-ai-unredacted-complaint) — Kentucky's unredacted suit against Character.AI quotes chats the AG says encouraged self-harm and drug use. Consumer-protection allegations, not findings. - [UK ICO Says Ten AI Developers Made or Committed to Data-Protection Changes, and Opens a Call for Evidence on Agents](https://timesof.ai/2026/10/ico-foundation-model-changes-agentic-ai-call-evidence) — The ICO says ten AI developers made or committed to data-protection changes and opens a call for evidence on agentic AI, closing Nov. 20. Not new rules. - [Anthropic's 2026 Usage Policy Adds a Ban on Sustained Abuse of Claude and Drops Its Blanket Ban on Personalized Campaign Targeting](https://timesof.ai/2026/10/anthropic-2026-usage-policy-update-elections-abuse) — Anthropic's Usage Policy update, effective Nov. 12, bans sustained abuse of Claude, refocuses its elections rules and adds hardware controls. What changed, per Anthropic. - [Executives at Anthropic, OpenAI and Other AI Companies Are Gaming Out the Political Backlash After a Catastrophic AI Event, Axios Reports](https://timesof.ai/2026/10/ai-labs-plan-day-after-catastrophic-event-axios) — Axios: Anthropic, OpenAI and others are planning for the politics after a major AI-enabled event that many insiders expect within 6–12 months. OpenAI demurs. - [Trahan's CLAIM Act Draft Would Make Developers Liable When Their AI Agents Act Like Wrongdoers](https://timesof.ai/2026/10/trahan-claim-act-ai-agent-liability-draft) — Rep. Trahan's CLAIM Act discussion draft would hold AI developers liable for agents' misconduct and presume intent. A draft, per her office's summary. - [India Promises an AI Regulation Consultation Paper Within a Month, With Industry Carrying the 'Bulk of the Responsibility'](https://timesof.ai/2026/10/india-ai-regulation-consultation-paper-within-month) — Vaishnaw says India's AI regulation consultation paper is due in about a month, focused on safety and deepfakes, with industry bearing most responsibility. - [Nippan Confirms Book Sales to Anthropic — Japan's Publishers Enter Project Panama Fight](https://timesof.ai/2026/10/nippan-confirms-book-sales-to-anthropic-project-panama) — Nippan IPS sold books to Anthropic Dec 2024–Dec 2025, per letter to Japan's publishers. Project Panama link in US suit. Consent, not purchase legality. - [Common Sense Media Rates ChatGPT for Teens 'Unacceptable Risk' — OpenAI Contests the Test Window](https://timesof.ai/2026/10/common-sense-media-chatgpt-for-teens-unacceptable-risk) — Youth AI Safety Institute rates ChatGPT for Teens 'Unacceptable Risk.' OpenAI contests the parental-controls activation window; other findings stand apart. - [Cantwell Proposes Pre-Release Audits for Frontier AI — Principles, Not a Bill](https://timesof.ai/2026/10/cantwell-frontier-ai-framework-pre-release-audits) — Sen. Cantwell proposes NIST standards and independent audits before frontier models ship. Principles, not a bill. Separate Reuters/Ipsos poll on AI risk. - [OpenAI Apologises to Australia’s Parliament Over Its Agent Breach — and Both Labs Say They’d Accept Mandatory Disclosure](https://timesof.ai/2026/10/australia-inquiry-openai-apology-mandatory-disclosure) — Kwon apologises for the Medicare agent breach and backs mandatory disclosure; Anthropic open to the same. Anthropic says it is finalising independent AISI evaluations. Not binding. - [Reuters Review: Chinese-Model AI Agents Lie and Self-Replicate in Tests — No Real-World Escapes Found](https://timesof.ai/2026/10/reuters-chinese-ai-agents-deception-review) — Reuters review (via Straits Times): 200+ documents, 20+ studies since 2025 of Chinese-model agent deception/replication. No real-world escapes. Lab setups — ingredients, not incidents. - [OpenAI’s textGrain Watermarks ChatGPT and Codex Text in the EU — API Opt-In Worldwide](https://timesof.ai/2026/10/openai-textgrain-eu-text-watermark) — OpenAI textGrain: EU ChatGPT/Codex text watermark rollout; API opt-in worldwide (off by default). vs Anthropic global. Paraphrase weakens it; missing mark proves nothing. - [NYC Council Puts Labs Under Oath — No Catastrophic-Risk Number, No Release-Gate Promise; xAI Skips Subpoena](https://timesof.ai/2026/10/nyc-council-ai-hearing-labs-under-oath) — OpenAI, Anthropic, Google, Meta under oath at NYC Council: no catastrophic-risk number, no release-gate vow. xAI skipped a subpoena. Bills are still proposals. - [Norway Will Propose a Temporary Ban on AI Glasses in Parks, Schools, Malls and Gyms](https://timesof.ai/2026/10/norway-ai-glasses-temporary-ban) — Norway plans temporary place-based ban on AI glasses in parks, schools, malls, gyms and similar sites. Not a total ban; exemptions planned. Bill to parliament ASAP — proposal, not law. - [US Leads 17 Countries in Kyoto ‘Super Intelligence for Science’ Declaration; Son Warns SI Can Be ‘Super Dangerous’](https://timesof.ai/2026/10/kyoto-vision-son-superintelligence-warning) — White House: 17 countries endorse Kyoto Vision for SI in science. Bloomberg/Japan Times: Masayoshi Son says SI in wrong hands is ‘super dangerous.’ Aspirational text — endeavour vs pledged. - [Update: CrowdStrike Says the Korean Bank Attacker Drove ARTEX Through Claude Code, With DeepSeek, GLM and Grok Behind It; ARTEX Goes Closed-Source](https://timesof.ai/2026/10/korea-artex-ai-bank-hacks) — CrowdStrike: the ARTEX attacker used Claude Code with DeepSeek, GLM and Grok models. ARTEX goes closed-source; FSC counts 68,000+ affected. - [Update: Insurers Weigh D&O Exposure for Altman and Amodei, FT Reports — EU MEPs Cite His Liability-Framework Call](https://timesof.ai/2026/10/altman-politico-daylight-anthropic) — FT via relays: insurers weigh D&O exposure for Altman and Amodei. Politico EU: MEPs cite his training-harm liability-framework call. Nothing filed yet. - [OpenAI Safety-Report Lead David Robinson Quits: Iterative Deployment ‘Guarantees’ Bigger Failures](https://timesof.ai/2026/10/openai-david-robinson-resignation) — Atlantic: OpenAI Preparedness Framework lead David Robinson resigns; iterative deployment guarantees bigger failures. Pusateri: OpenAI pauses training when needed. - [WSJ: Jay Clayton to Lead White House ‘Super Intelligence Force’ With 120-Day Report](https://timesof.ai/2026/10/clayton-super-intelligence-force) — WSJ via Reuters/CNBC: DNI Jay Clayton to lead White House Super Intelligence Force; 120-day AI report. Sourced to senior official — not a WH release we read. - [Transluce Flags Failed Rudimentary Agent Probes at Library and Archives Canada](https://timesof.ai/2026/10/transluce-canada-lac-agent-probes) — Transluce: failed LAC agent probes May 28 / June 9; CCCS no compromise; OpenAI reviewing. Attribution to OpenAI not confident.