AI Safety Summit Round-up: Part Three
By Fay Capstick
In the past two weeks, we have taken a deep dive into the recent AI Safety Summit, held and hosted by the UK. This week we will wrap this up by asking what else has happened. We will look at Grok (another offering from Mr Musk), and the huge new UK investment in supercomputers, and we ask what this means for our industry.
Anything else happened recently with AI?
Yes! Grok!
Even after all that has been learned and discussed about AI from the summit, things might still go wrong with the way that AI is developed and integrated into our world.
Just a few short days after attending the summit, Elon Musk announced that his company, xAI, had created a new AI chatbot called Grok, in only 8 weeks. Worryingly he declared it to be powerful and that it would have fewer guardrails around it. By this, he means that it is designed to answer questions with more humour and sarcasm. However, it will also answer questions that other AI systems are programmed to skip. This could include explicit, violent, or illegal content, and it hardly seems a sensible way to go. A request for clarification from Wired was met with an autoresponder stating they were ‘busy now’.
More UK Government investment in supercomputers
The government has announced that it will be investing in Bristol’s “Isambard-AI”, and also in a new supercomputer, “Dawn”, at Cambridge. This investment will boost UK access to supercomputing power by 30x.
The two computers will form the government’s ‘AI Research Resource’, with an investment of £300 million (£225 million going to the Bristol computer). This will help researchers from the AI Safety Institute, as they will need supercomputer access to undertake their AI testing. They will be getting priority access to these machines. It is expected these two new supercomputers will be running by summer.
The supercomputer at the University of Bristol, ‘Isambard-AI’ will become Britain’s most advanced supercomputer, at 10x the speed of the current most advanced. This will be due to ‘5,000 advanced AI chips from Nvidia in a supercomputer built by Hewlett Packard Enterprises (HPE). It will have a speed of 200 petaflops.
The other supercomputer, ‘Dawn’, is a result of a partnership between Dell and StackHPC (a UK SME). Dawn will be using water cooling to help cool its 1,000 Intel chips. We have seen before how data centres use huge amounts of power, and any methods to reduce this environmental impact are positive, though, of course, water is a limited resource too. ‘Dawn’ should be running in the next couple of months.
Does any of this help with the skills gap?
Absolutely, the government have announced a £118 million skills package as they recognise the problem that the UK faces with a lack of workers with the necessary skills. There will be funding for 12 centres for Doctoral Training in AI (locally to Parker Shaw in Southampton, at the University of Southampton, there will be the UKRI AI Centre for Doctoral Training in AI for Sustainability).
In a wider effort to plug the skills gap, there will be a new visa scheme to bring workers with the skills needed into the UK. Students at all levels will be encouraged to take AI and data courses. The overall result should be the UK generating as well as encouraging the top talent to be in the UK.
What do we think?
After all the talk at the summit about testing and regulation and sandboxing, this seems a reckless move on the part of xAI and Mr Musk. It feels as though Musk fully appreciates the risk of AI, but feels that anything he produces can’t or won’t be part of the problem (an idea we mentioned in a previous blog, that it can’t possibly be your child who is the bully). This seems extremely shortsighted and arrogant of Musk. He is exhibiting behaviour that, if it continues, could be damaging to the whole of humanity.
Frontier AI and how we proceed can be considered the most urgent challenge facing our planet due to the potential for it to go wrong. There really is nothing more important. If we can get a handle in place on how to regulate AI (and regulation is essential), then humanity can start to enjoy the benefits that AI can bring us in terms of medical, social, and environmental benefits. The possibilities are limitless and life-changing, both if we get it right and if we get it wrong. It is beyond exciting that we are witnessing what will be a pivotal point in the history of humanity.
Final thoughts
At Parker Shaw, we have been at the forefront of the sector we serve, IT & Digital Recruitment and Consulting, for over 30 years. We can advise you on all your hiring needs.
If you are looking for your next job in the IT sector please check our Jobs Board for our current live vacancies at https://parkershaw.co.uk/jobs-board