Ctrl+AI+Reg - 5 December 2025
Your shortcut to AI regulation, law and policy updates around the world.
AI Regulation Updates
In this issue:
Updates from Canada, US, EU, UK, South Korea.
Opinion on the “three camps” of China’s AI industry.
See more on my Global AI Regulation Tracker (English version | Chinese version)
Americas
🇨🇦 [3 December 2025] Canada releases world’s first standard for accessible, inclusive AI design: Accessibility Standards Canada has released the CAN-ASC-6.2:2025 standard, titled “Accessible and Equitable Artificial Intelligence Systems,” which is described as the world’s first standard on accessible and equitable AI. The standard emphasises that AI systems should provide comparable benefits to people with disabilities, avoid causing them disproportionate harm, protect their rights and freedoms, and involve them in all stages of the AI lifecycle. It outlines requirements for organisations to implement processes that support accessible and equitable AI governance, planning, procurement, development, deployment, and monitoring. Additionally, the standard mandates accessible education and training on AI, involving people with disabilities in the creation and delivery of such programs, to promote AI literacy and enable informed participation.
🇺🇸 [3 December 2025] Republicans drop Trump-ordered block on state AI laws from defense bill: It is reported that a recent attempt to include a ban on state-level AI regulation within an annual defence bill - the National Defense Authorization Act (NDAA) - has been rejected due to bipartisan opposition. House Majority Leader Steve Scalise indicated that Republican leaders, with President Trump’s support, plan to seek alternative avenues to introduce this measure. This follows a previous unsuccessful effort to impose a 10-year moratorium on state AI laws, which faced strong resistance from both parties. Scalise acknowledged that the NDAA was not the appropriate vehicle for such a provision and echoed President Trump’s suggestion to introduce the ban as a separate bill. A leaked draft executive order indicates that Trump is considering unilateral action, though these efforts are currently on hold.
🇺🇸 [3 December 2025] Nvidia CEO talks chip restrictions with Trump: It is reported that Nvidia CEO Jensen Huang met with President Donald Trump to discuss US export controls on advanced AI chips, expressing concerns that these restrictions have significantly reduced Nvidia’s market share in China. In the background, it is also reported that US lawmakers are considering to not include the GAIN AI Act in the annual defense package, known as the National Defense Authorization Act. The GAIN AI Act would have required chipmakers like Nvidia and Advanced Micro Devices to give US companies first pick on their AI chips before selling them in countries like China and other adversary nations (effectively limiting the company’s ability to sell chips to such markets).
🇺🇸 [2 December 2025] AI Civil Rights Act to reintroduced: US Senator Edward J. Markey (D-Mass.) and Representative Yvette Clarke (NY-09) have reintroduced the AI Civil Rights Act, aiming to prevent the use of biased and discriminatory AI algorithms in critical decision-making processes affecting American individuals. It has three key provisions: (1) to prohibit companies from offering, licensing, or using algorithms that discriminate on the basis of protected characteristics, like race or sex; (2) to require AI companies to conduct independent audits to identify and mitigate algorithmic discrimination; and (3) to give individuals the right to appeal an AI-based decision to a human, and empower the Federal Trade Commission, state attorneys general, and private individuals to enforce this part of the law. The bill was previously introduced in the US Senate in September 2024 but stalled in the 118th Congress session and expired thereafter.