Policy & regulation

Anthropic immediately suspended all user access to Claude Fable 5, citing a US government directive. Other Claude models are unaffected. New chats default to the user's chosen model or Opus 4.8, and existing Fable 5 sessions throw errors. API calls also fail, with migration to other Claude models recommended. The post does not disclose the directive's specifics or a timeline for reinstatement.

Amazon shared its global data center water consumption for the first time: 2.5 billion gallons in 2025, at 0.12 liters per kilowatt-hour. That's a 2% drop from 2024 even as operations expanded. Amazon also claims better water efficiency than some Big Tech rivals, pointing to a chart in its report. The disclosure came right after Seattle enacted a data center moratorium that some Amazon employees had pushed for. The post doesn't name the rivals or show the full comparison data.

A group of independent musicians is suing Google, claiming it trained its Lyria music model on YouTube creators' uploads. Google's response is evasive: it says the terms of service allow it, but refuses to confirm or deny whether it actually did. The post doesn't disclose Lyria's training data sources or scale.

On June 10, China's MIIT issued an implementation opinion on 'AI + information communications.' It orders accelerating backbone network upgrades to 400Gbps/800Gbps between eastern, central, and western hub nodes, and deploying 400G metro networks with all-optical cross-connect to achieve millisecond-level latency for computing access. It also pushes AI-driven network tech: 5G-A/6G air interface intelligence, self-optimizing networks, space-based computing networks, and agent internet. Operators are encouraged to use AI for smart home assistants, 3D viewing, and other consumer services. The post does not disclose investment figures or a timeline.

The European Commission issued an interim measure on June 9, ordering Meta to give third-party general-purpose AI assistants free access to WhatsApp until its antitrust investigation concludes. Meta banned external AI from the WhatsApp for Business API in October 2025, then switched to a paid model in March 2026. The Commission sees the paid access as effectively continuing the ban and harming smaller competitors.

A German district court ruled Google is directly liable for AI Overviews content after one overview wrongly linked two publishers to fraud, and the cited linked sources did not contain the statements.

China plans to invest about 2 trillion yuan, or $295 billion, over five years to build nationwide data centers, with funding covering large-scale data center infrastructure for domestic AI development.

The Pentagon reinstated Alibaba, Baidu, and BYD as U.S. national security risks after removing the three companies in February; the RSS snippet does not disclose the legal basis, timetable, or company responses.

China’s MIIT and SASAC require key products including humanoid robots to complete application validation and enter routine deployment by the end of 2026, targeting more than 100 high-value scenarios and deployment at the ten-thousand-unit scale.

OpenAI said on Monday it has entered its third phase, naming three goals: automated AI researchers, faster economic growth, and personal AGI for everyone, while calling for an international body to manage AI risks.

A judge blocked Trump’s $100,000 H-1B visa fee, ruling that the six-figure charge was an unauthorized tax and ordering it suspended; the RSS snippet does not disclose the court, case name, or implementation timeline.

FT says AI can use CCTV data to identify targets, and Russia paused a surveillance system after the killing of Iran’s Supreme Leader; the RSS snippet does not disclose the system name, model mechanism, vendor, or timeline.

OpenAI outlined its third-phase plan with three goals: build an automated AI researcher, accelerate the economy, and give every person a personal AGI. Sam Altman and Jakub Pachocki said OpenAI internally believes AI systems may perform a significant fraction of its research by March 2028, while alignment, safety standards, and international coordination remain explicit conditions.

Bernie Sanders met with Sam Altman to discuss transferring 50% ownership of major U.S. AI companies to the public, while the article cites a Quinnipiac poll saying 80% of Americans are concerned about AI.

U.S. House lawmakers released a draft AI bill that would bar individual states from setting their own AI regulations and concentrate regulatory authority at the federal level; the RSS snippet does not disclose the sponsors, committee path, voting schedule, or the federal mechanism that would replace state-level rules.

Trump said he has discussed deals with AI companies that let Americans benefit from AI’s success, and CNBC reported the administration is discussing an OpenAI equity stake; the article does not disclose the stake size or deal terms.

The title says Trump signaled interest in the US owning stakes in top AI labs; the body is a Bloomberg 403 anti-bot page and does not disclose the target labs, stake size, policy mechanism, or timing.

Police in England and Wales were told to halt AI use in court statements until safeguards are in place; the RSS snippet cites the head of Police.AI but does not disclose the specific safeguards or enforcement mechanism.

Trump said the US may take equity stakes in AI companies, but the FT article body is a subscription page and does not disclose stake size, target companies, transaction terms, or policy mechanism.

The Intercept says the Pentagon runs an AI propaganda mill targeting Latin America; the snippet discloses the target region, AI-generated content distribution, and 100 Hacker News points, but not budget, vendors, model stack, or operational timeline.