BBC 5 Live – Seoul Safety Summit
BBC5 Live Seoul Safety Summit
SPEAKERS
BBC, Aarti Samani, Victoria Jenkins
BBC
Talking of artificial intelligence today, the balance between the risks and the opportunities, as Victoria’s touched upon a bit earlier in the program. It’s going to be the focus of a big summit over the next couple of days. Rishi Sunak is going to be there. He’s going to be at the AI Seoul Summit in South Korea. He’s going to say that artificial intelligence will create major opportunities for human progress, but managing its risks is one of the most profound responsibilities faced by world leaders. Just six months after we had the AI Safety Summit at our very own Bletchley Park, the technology though, including chat bots like ChatGPT, Gemini, that’s Google’s one, if you’re keeping up with it, Microsoft’s copilot, have seen rapid growth in that time. Aarti Samani is Chief Executive of the AI consulting firm, Shreem Growth Partners, and joins me now. Aarti. Hello,
Aarti Samani
Hi. Good morning
BBC
Thanks so much for being with us, Aarti. Summit, after summit after Summit. What is all the conversation leading to, at the moment, with those at the forefront of the industry. What are people trying to achieve?
Aarti Samani
Yeah, great question. You’re right, summit after summit. So we’re six months into the AI Safety Summit, the first one held in Britain in November. At that time, the commitment was to create a report to identify, or at least help gain understanding of what the potential risks are of AI, and the rate at which AI is moving, what risks are we generating. Within six months, Britain, in collaboration with other nations, has produced a report, which was released yesterday. And the report is called International Scientific Report on the safety of advanced AI. It’s an interim report. That’s what will be discussed in this summit, which is being held in Seoul today and tomorrow. And the idea is to really understand the risks and then to get a collective global way forward on mitigating some of those risks. So the report we saw highlights a whole bunch of things, including what is called frontier risks, which is risks of the unknown unknown. So what can potentially happen, including how humanity can be destroyed by AI. Now that was the topic of conversation in the last summit.
BBC
I love it, how you casually mentioned that just chuck that one in our team, including, on a long list, just how humanity can be destroyed by AI!
Aarti Samani
Yeah well, and to be honest, yes, humanity may be destroyed at some point. Who knows? It’s not an immediate threat that concerns me or concerns lay people on the street, right? But my concern is, what are we doing about risks like deepfakes, which is affecting businesses, it’s affecting individuals on a daily basis, and thankfully, the report highlights that as a risk. So, what I would like to see coming out of the summit from Seoul is what government, public sector and policy makers are doing collectively to mitigate risks like deepfakes, like synthetic media. We saw Arup, a UK business targeted by deepfakes earlier this year.
BBC
Yeah, the CFO being faked and a lot of employees falling for it. Aarti, just hold there a sec, because Victoria was chatting a little bit earlier. Victoria Jenkins, who’s the Chief Executive of Unhidden, a fashion brand. Victoria, I just wonder what you might say to Aarti, or those going to the summit today, touching on a little bit of what we discussed earlier, about what should be high up the agenda when discussing where AI goes next.
Victoria Jenkins
I think, you know, as we’ve just been hearing, I think there needs to be some kind of rules and policy around protecting businesses, protecting people because, you know, the deepfakes are the problems with all of what it could cause. It’s just another way of removing, you know, people can lose, literally lose out on money, it could cost them their business if somebody else is impersonating them or ripping it off. So I think there needs to be a lot of regulation around what it what it can do to make it safer.
BBC
Aarti in the world of for the output of this AI stuff. We were getting into this a little bit earlier as well, and concerns that the output isn’t necessarily fair. We were talking about people with disabilities earlier. You could go across many different minorities and see concerns that AI is not working for them. How high up the agenda is that? It’s like we’re racing ahead to see what this technology can possibly do. But hang on, are we getting the first bits right to ensure that the data that’s being used is, is fair in itself and then being used in a fair way?
Aarti Samani
Yeah, that’s a great point. So bias in AI is very very topical, and thankfully, that is one thing that’s being looked at quite seriously by everyone in terms of the data that is used to train the algorithms, which ultimately generate the output. Unfortunately, AI is a reflection of the society that we live in, so there is bias in the data itself, because we do not have, all the data from all segments or groups of the society representative. But also, the good thing about AI is that it allows us to create synthetic data, which means given what we know of the world, using AI, we can simulate a data set which is representative of the world, even though that data does not exist. So that is being looked into. Now, I’m not saying that it’s there, that it’s all very good yet, but it is something that technologists, tool creators, etc, are working on it. And there’s a big drive from the government on it also.
BBC
Aarti, thank you so much for joining us, talking to us about the array of summits going on at the moment. Aarti Samani, Chief Executive of that AI consulting firm, Shreem Growth Partners.
KEYWORDS
AI risks, AI opportunities, AI Seoul Summit, AI safety, frontier risks, deepfakes, synthetic media, AI bias, synthetic data, AI regulation, AI fairness