Jump to content

AI Law Wiki News for September 14, 2026

From AI Law Wiki
Revision as of 16:34, 14 September 2026 by AILawWikiAdmin (talk | contribs) (Add Microsoft model-code consultation and September 10 Amazon rehearing coverage)

AI-generated text. This page was generated using artificial intelligence.

China responds to Amodei on AI cooperation and export controls

On September 14, 2026, China’s foreign ministry urged international cooperation on AI in response to Anthropic chief executive Dario Amodei’s proposal to retain US restrictions on advanced chips and chipmaking equipment. Spokesperson Guo Jiakun said confrontation and fear-based competition would undermine global AI governance.[1]

Amodei’s essay advocates preserving a US technological lead while slowing frontier development. It also calls for cooperation with China, emphasizing verification of any agreement. These are proposals, rather than an agreed international regime.[2]

See China policy on catastrophic AI risk and United States policy on catastrophic AI risk for the broader policy frameworks.

UK parliamentary committee calls for AI bill and independent oversight

On September 14, 2026, the UK Parliament’s Joint Committee on Human Rights called for a dedicated AI bill and an independent statutory oversight body. It recommended stronger obligations for higher-risk systems across development and deployment, including prior approval for systems posing high human-rights risks. These are committee recommendations, not enacted requirements.[3][4]

The report forms part of the debate over UK AI oversight and prevention of catastrophic harm, alongside broader discrimination, privacy and remedy concerns. Responding to questions on BBC Breakfast, education minister Georgia Gould pointed to existing action on sexualised deepfakes and cyber security and said the government was monitoring AI’s development; this did not announce adoption of the committee’s proposed regime.[4]

Microsoft opens consultation on human-control code for its AI models

On September 14, 2026, Microsoft AI opened a six-week consultation on a draft code for its own MAI models. The proposed training and evaluation standards call for models to accept human correction and shutdown, stay within authorized goals, and remain understandable to auditors. Microsoft plans to publish a revised version later in 2026.[5]

Microsoft AI chief Mustafa Suleyman told Reuters that the revised code would be used to train the company’s models. This is a proposed company governance framework, not legislation or proof that deployed models already meet every standard.[6] See United States policy on catastrophic AI risk for related approaches to human control.

References