[Documentation] AI Policy and Regulation Updates

[seanmcilroy29]

AI Policy and Regulation Updates

• In California, the state Senate pass the AI Safety and Innovation Bill, which open-source advocates have challenged as potentially damaging to open-source use cases of AI. You can read the Twitter comments to get a flavour of it. The second link is an analysis by a big law firm that is useful in understanding the bill. It fails on at least two levels. It sets an arbitrary training FLOPs definition of frontier models and also includes models that can match the performance of state-of-the-art models, which, for specific tuned tasks, covers a huge chunk of open source models today — Link, Link

• GitHub proposes changes to its Acceptable Use Policy in response to deep fakes and new imperatives of responsible AI — Link

• Former OpenAI BOD member Helen Toner pens a strong op-ed saying AI firms must be regulated by governments more tightly — Link Governor of the state of Colorado in the USA signs extensive AI regulation bill — Link, Link

• Can California fill the U.S. Federal void on AI regulation? — Link

• California Governor Newsom expresses concern about AI regulations smothering innovation — Link

• How India is viewing the OpenAI drama — Link

• OpenAI published an update to its safety policies, including ten safety practices and the establishment of a Safety and Security Committee — Link

• Meta has offered to limit the way it uses data from advertisers on its Facebook property in response to United Kingdom concerns about privacy — Link

[tholewis]

Thank you for all the great links here! Lots of good insight across this space. 🌳