PolicyDaily digest, November 12, 2026
Policy Shifts Shape AI's Future
This week brings major changes to how governments and companies approach AI regulation, from new safety requirements in Australia to ongoing US efforts to control chip exports to China. Congressional investigations into Nvidia's practices and evidence of export control circumvention highlight growing tensions around AI chip access and national security.
Chinese AI Models Show Signs of Using Restricted Nvidia Chips
Forensic analysis discovered that six open-weight AI models from Chinese companies like DeepSeek were likely trained using export-controlled Nvidia Blackwell GPUs, despite US restrictions on these sales. The findings suggest Chinese firms may be successfully circumventing American export controls on advanced semiconductors.
Why it matters: This evidence suggests Chinese companies are circumventing US export controls on advanced chips, challenging the effectiveness of the Biden and Trump administrations' national security restrictions on AI chip sales.
Sources: via Techmeme · via Techmeme · via Techmeme
Australia Requires Independent Safety Testing for Advanced AI Models
Australia's government proposed new legislation that would mandate independent third-party audits of frontier AI models, with penalties for companies that fail to comply. This represents one of the first national laws requiring outside verification of advanced AI system safety.
Why it matters: This represents one of the first national laws requiring independent auditing of frontier AI models, setting a precedent for how governments can regulate advanced AI development.
Sources: The Decoder
Commerce Secretary Delays Enforcement Against Chip Diversion Suspects
Commerce Secretary Howard Lutnick reportedly slowed action against companies suspected of illegally sending Nvidia AI chips to China, despite the administration approving export restrictions for a third time this year. The enforcement delays raise questions about how strictly the administration will pursue its stated goal of preventing advanced chip sales to China.
Why it matters: Restrictions on AI chip exports to China affect global AI development competition and U.S. national security policy around AI infrastructure access.
Sources: via Techmeme · MIT Technology Review
OpenAI Funds $2 Million Child Safety Research Initiative
OpenAI committed $2 million to research measuring risks and opportunities of AI for children's wellbeing. The investment reflects growing industry concern about AI's effects on young people.
Why it matters: Major AI labs investing in child safety research signals growing concern about AI's impact on young people and commitment to developing shared measurement tools for this critical issue.
Sources: OpenAI
Nonprofits Sue Tech Giants to Slow Advanced AI Development
Five nonprofits filed a lawsuit in San Francisco against major companies including Anthropic and Google, seeking a court order to block development of more advanced self-improving AI models. If successful, the lawsuit could set legal precedent requiring companies to justify major capability advances to courts.
Why it matters: If successful, this lawsuit could create legal precedent constraining AI development timelines and force companies to justify model capability advances to courts, potentially slowing the AI arms race.
Sources: Politico
Nebius Plans $40 Billion AI Infrastructure Expansion
Nebius is pursuing a massive $40 billion expansion of AI computing infrastructure, funded through customer prepayments and loans. The financing approach raises questions about how AI infrastructure buildouts will be funded and whether companies can sustain such ambitious expansion plans.
Why it matters: How AI infrastructure companies finance massive buildouts affects competition, costs, and availability of computing resources for AI development.
Sources: TechCrunch