AI safety and regulation news covering governance frameworks, AI legislation, safety research, alignment, and policy debates on artificial intelligence.
The Trump administration is weighing plans to create an independent regulator to vet AI model safety, modeled after the Financial Industry Regulatory Authority (Finra). Treasury Secretary Scott Bessent helped develop the proposal, which would report to the SEC. The plan is under review by White House Chief of Staff Susie Wiles.
At the Vatican-hosted Global Nobel Laureates Assembly, experts from OpenAI, Google DeepMind, and others signed the Rome Declaration for an Unarmed and Disarming Peace, addressing AI and nuclear war risks. The declaration builds on Pope Leo XIV's encyclical on protecting humans in the age of AI.
Former Air Force Secretary Frank Kendall argued that AI should be allowed to make lethal decisions when it outperforms humans, provided rigorous testing and verification ensure reliability. He stressed training operators to use AI tools properly.
OpenAI will alert parents when a teen's ChatGPT account is deactivated for violating policies on violent threats. The feature, developed with monitoring firm Moonshot, also adds Study Mode hints and break reminders for teens.
At the World AI Conference in Shanghai, delegates warned that fragmentation of AI governance risks isolating emerging markets and concentrating advanced computing in wealthy nations. Tsinghua scholar Xue Lan called for international organizations to improve interoperability between national systems, while former Mongolian President Enkhbayar cautioned against stifling early-stage progress in developing countries.
A bipartisan House bill would require human oversight for lethal AI weapons use by the US military. Commanders must verify AI targets with non-AI sources for five years. Missile defense is exempt. The bill codifies existing Pentagon policy into law.
Public First Action, a 501(c)(4) co-founded by former Rep. Chris Stewart, launched the Conservative AI Consensus Project with $15 million to support Republican lawmakers who back AI guardrails. The project starts with over $7 million in ads featuring four GOP members of Congress and will fund national polling on conservative AI attitudes.
San Francisco Mayor Daniel Lurie urged California regulators to strengthen autonomous vehicle rules after Waymo robotaxis were immobilized during July 4 celebrations and a December 2025 power outage. His proposal requires AV operators to demonstrate operational readiness for major events, including quick vehicle removal and real-time data sharing with local agencies.
This commentary argues that the US government must act swiftly to regulate AI trade secrets to maintain competitive advantage, while economists urge immediate policy intervention. The piece suggests that current inaction risks economic and security consequences, framing the debate as urgent and requiring coordinated federal response.
California regulators confirmed that existing anti-discrimination laws apply to AI hiring tools, and lawmakers are advancing new transparency and testing requirements. Stanford research found racial disparities in a leading third-party hiring platform, disadvantaging Black and Asian applicants.