the summit is a follow—up to the summit in november at bletchley park, and its aim essentially is tothat work. it is a two—day summit, the start of it virtually attended by prime minister rishi sunak and his south korean counterpart. the second day brings together digital ministers, and they will look at a range of issues including things like the environmental impact of ai, issues of bias, the impact onjobs, which is interesting because the focus on the bletchley park summit was the catastrophic scenarios. those particular items on the agenda much more here and now concerns. broadly, thatis more here and now concerns. broadly, that is what the summit is hoping to achieve. there was talk about it possibly being lacklustre and so much surprise when this morning we get this announcement of 16 firms coming together with this agreement to check out their systems and to have red lines, where if they think frontier system is approaching something seriously risky, taking action including potentially stopping development. it is action including potentially stopping development. it is always

of its kind, it is co—hosted with britain after the inaugural event in november last year at bletchley parkndon. francine bennett is interim director at the ada lovelace institute — an independent research institute specialising in al — and is in south korea for this summit. a very warm welcome to the programme. so first of all, what will be achieved in seoul, do you think?— do you think? well, that's a aood do you think? well, that's a good question. _ do you think? well, that's a good question. i _ do you think? well, that's a good question. i hope - do you think? well, that's a good question. i hope what| do you think? well, that's a - good question. i hope what will be achieved in seoul is moving on from the conversation from the bletchley summit last year. so we had some really good initial conversations there, and i think what wasn't to see is shaping the development of this, potentially very new technologies, but in the interests of people and society. so you mentioned our research institute is an ai research institute is an ai research institute. actually, we research how to make ai

it was launched last november with a conference at bletchley park — the wartime code—breaking centreal safety. the bbc technology editor, zoe kleinman, told me more. the conversation in the last six months has massively changed, i think, on the subject of safety. if you remember the discussions back then, it was real kind of terminator, doomsday scenarios — that has kind of gone away now. it's got a much broaderfocus. i've just read a report that's come from the ai safety institute in which it says, "actually, we think ai is probably less likely "to create a lethal bio—weapon "or orchestrate a sophisticated cyber attack. "and we kind of don't really think that humans are in "any danger as yet of losing control of ai and it doing, "you know, something horrendous to civilisation". but is it less dangerous than it was? surely not. we heard a lot, in november, the phrase "marking your own homework", and it does still feel a little bit like that is what's happening. and bear in mind, everything that's been agreed both at bletchley and in seoul is voluntary. and i think when we see more le

developers signed an international agreement that builds on the commitments secured last year at bletchley parkmpanies from china, america and the middle east have agreed not to develop or deploy ai models that cannot maintain risk below a certain threshold. eu legislation, a new international agreement — but where is the overlap with china and the united states, two of the biggest developers? and where do the big powers diverge? we will get the thoughts tonight of miles taylor, who is advising the us congress. with me in the studio, our regular ai contributor and author on technology, stephanie hare. hello. can we talk about the uk and the south korea summit? the british government setting out this new agreement this week, rishi sunak said it's a world first to have so many leading ai companies all agreeing to the same commitments on al safety. how do you think this is ai safety. how do you think this is significant? it ai safety. how do you think this is significant?— ai safety. how do you think this is siunificant? , , , , significant? it depends, it depends because it is _ significant? it d

developers signed an international agreement that builds on the commitments secured last year at bletchley parknd west ham midfielder lucas paqueta has been charged with misconduct by the football association in relation to alleged betting rule breaches. the charges relate to four games over an 18 month period, where it's alleged paqueta deliberately tried to get a card from the referee in order to influence betting markets. the fa investigation was triggered by suspicious betting patterns. he has until the 3rd ofjune to provide a response, although he's already publicly denied any wrongdoing, on social media saying... "i will fight with every breath to clear my name." here's our senior football reporter simon stone. what they have said, as you say, in their statement is basically that he is intentionally seeking to receive a card for the purposes of effecting the betting markets in those games in order for one the betting markets in those games in orderfor one or more the betting markets in those games in order for one or more persons to profit from betting

yes, well, it comes after a successful summit at bletchley park last year. yeah >> well, the chairmanashik parmar , joins us now. what do parmar, joins us now. what do you make of it? rashik i think it's brilliant that we're actually making progress here. >> you know, a! is a part of everybody's life now. when you when you go look at somewhere like there's an ai when you go look at somewhere like there's an a! for that, there's, there's 12,000, a! tools out there, supporting 15,000 jobs. it's got tremendous opportunity to make a positive impact on society and help solve some of those societal issues. but of course, i mean , the whole but of course, i mean, the whole point of this sort of seminar or whatever it is this meeting, is to try and iron out what the dangers are, and then, i suppose, try and get everybody to agree , but that's never going to agree, but that's never going to agree, but that's never going to be possible. is it? i mean, inventors, code writers , people inventors, code writers, people who specialise in these things are always going to be one ahead, going to want t

it was launched last november with a conference at bletchley park — the wartime code—breaking centremuch broaderfocus. i've just read a report that's come from the ai safety institute in which it says, "actually, we think ai is probably less likely "to create a lethal bio—weapon "or orchestrate a sophisticated cyber attack. "and we kind of don't really think that humans are in "any danger as yet of losing control of ai and it doing, "you know, something horrendous to civilisation". but is it less dangerous than it was? surely not. we heard a lot, in november, the phrase "marking your own homework", and it does still feel a little bit like that is what's happening. and bear in mind, everything that's been agreed both at bletchley and in seoul is voluntary. and i think when we see more legislation come in — we'vejust seen now the eu ai act come in — there you are, that's law now, these companies have got to abide by european law. and i wonder whether this goodwill that we've seen so far between policy—makers and the industry might become slightly less amicable once they're forced to do