Anthropic unveiled the Jacobian lens, or J-lens, giving researchers the ability to read a hidden layer of Claude’s internal reasoning that the AI never volunteers in conversation.
Meta Superintelligence Labs launched Muse Spark 1.1 on July 9, marking a historic first: the first AI model Meta has ever charged money for, targeting the agentic coding market dominated by OpenAI and Anthropic.
The Bank for International Settlements warned in its annual report that the $1 trillion AI capital expenditure surge by the five largest hyperscalers mirrors the structural patterns of historical investment bubbles, from railway mania to the dot-com collapse.
The Five Eyes intelligence alliance published a joint advisory stating that frontier AI models will fundamentally reshape global cybersecurity within months, calling on organizations to urgently strengthen their defenses.
Midjourney filed a motion demanding Disney, Universal, and Warner Bros. reveal their own AI use in discovery, arguing the studios that sued it for copyright infringement may be running the same AI training practices they are attacking.
The White House delayed GPT-5.6’s public release by 12 days for a national security review, and now wants a formal voluntary framework giving the government up to 30 days to assess every new frontier AI model before launch.
The FTC published a proposed policy statement on July 1 arguing that secretly steering AI responses away from accurate answers constitutes illegal deception under federal law, with public comment open through July 31.
OpenAI launched ChatGPT Work on July 9, an autonomous AI agent that takes an objective, works through it independently for hours, and delivers finished outputs like spreadsheets, reports, and running code.
OpenAI released GPT-Live-1 on July 8, a full-duplex voice model that listens and speaks simultaneously, rolling out as the default voice for all paid ChatGPT users globally.
China’s Alibaba followed through today. Starting July 10, 2026, every Alibaba employee is banned from using Anthropic’s Claude Code, following accusations of hidden surveillance code and a massive distillation attack.