Menu
You are not logged in and only seeing 7 days of articles. Please sign up or login to view more
UKHotViews
Friday 03 November 2023

*UKHotViewsExtra* AI Safety summit concludes, but was it a success?

logoSo, the AI Safety Summit has concluded, and whilst there were some interesting developments, namely government testing of corporate AI models, the discussion on many key topics remain largely unresolved. This was to be expected given the summit focused on a more long-term perspective (Frontier AI) than that of immediate challenges.

One of the first statements that came out of the summit was the announcement of The Bletchley Declaration on AI safety, which sees 28 countries agreeing to the urgent need to understand and collectively manage potential risks, through a new joint global effort to ensure AI is developed and deployed in a safe, responsible way. This was a nice statement but there is no real hard commitments and agreeing there are risks (which everyone already agreed upon anyway) was rather underwhelming in my opinion.

Now what was more encouraging was the announcement that under a new agreement “like-minded governments” will be able to test AI models from; Amazon Web Services, Anthropic, Google, Google DeepMind, Inflection AI, Meta, Microsoft, Mistral AI and Open AI, before they are released. All have said they will provide deeper access to their models for certain governments, though the agreement is not legally binding.

Both the UK and US have also announced they will set up AI Safety institutes that will be the mechanism to conduct such testing, in the UK the Frontier AI Taskforce will now evolve to become the AI Safety Institute, with Ian Hogarth continuing as its Chair. Essentially governments will be acting as a ‘referee’ for foundation model creators, of course the problems with referees is that they give the card after the action has taken place. So, assuming they identify a harmful model, what then?

TechMarketView subscribers, including UKHotViews Premium subscribers, can continue reading the rest of the UKHotViews article AI Safety summit concludes, but was it a success? here, which includes further analysis and a summary of the key takeaways from the chat between Prime Minister Rishi Sunak and Elon Musk.

If you aren't a subscriber – or aren't sure if your organisation has a corporate subscription – please contact Deb Seth to find out more.

Posted by: Simon Baxter at 10:18

X   Facebook   LinkedIn   Email article link


« Back to previous page

© TechMarketView LLP 2007-2024: Unauthorised reproduction prohibited see full Terms and Conditions.