A BBC World Report interview (April 2024) with Tim Davies, CEO of Project Rising, about a new US–UK memorandum of understanding to jointly develop AI model safety testing. The discussion highlights concerns about rapidly evolving, next-generation AI models (noting current public-facing models like GPT-3.5/GPT-4 Turbo are ~18 months old) and stresses that businesses must adopt AI to stay competitive.
United States and Britain have announced a new partnership on the science of artificial intelligence safety. Amid growing concerns about upcoming next generation versions. Commerce Secretary Gina Raimondo and the British Technology Secretary Michelle Donilon signed a memorandum of understanding in Washington to jointly develop advanced AI model testing following commitments announced at an AI Safety Summit in Bletchley park in November. You remember that. Let's discuss this further. Joining me now is Tim Davis, CEO of AI consultancy Project Rising. Hello there, Tim. So the safety concerns, I wonder if you could just top line them for us. What are they?
So there are a large number of concerns around how these models could perform in the future. We're dealing with the development of a new form of intelligence and obviously with any new intelligence there are risks. So these safety tests evaluate whether the models are safe and will perform reliably in the functions that they're deemed to work in. So these tests, working jointly with the US and the uk, are to align and make sure that there is commonality in testing and approach.
Okay, so next generation versions. I mean, many people are only just catching up with the idea of AI. Where are we?
We're moving very, very quickly. So the primary model that most people are used to experiencing is within ChatGPT, and that's GPT 3.5 Turbo or GPT 4 Turbo. And these models are quite old now by relative standards, 18 months old. And there are new capabilities coming online all of the time that are of a similar level of quality. We'll shortly see another significant lift in capability when OpenAI, Anthropic and Google release their latest models.
I suppose it's really important, isn't it, for not only for businesses, but for governments to get on top of this. For the uk, collaboration with the US must be fantastic because the US are very much in front with this. How much is AI worth to a business? Why is it key that businesses are aware of it?
I think it's as significant as change or die. Businesses that adopt artificial intelligence will outperform businesses that don't. So those that are working with this technology will be able to move faster and deliver more and have a higher level of profit than their competition.
Okay, Tim Davies, CEO of AI consultancy Project Rising. It's a pleasure. Thank you.
