One in three people believe AI will eventually contribute to a mass mortality event

0:00Dave Roberts here. There you are surrounded by fans, sharing wings, sharing drinks, high-fiving random strangers. Everybody remembers the game. Nobody remembers the guy coughing behind you until a few days later. At 2 a.m., you wake up with a fever and your throats on fire. Now what? Urgent care? Close. ER? Slam. Telehealth? Maybe. But the pharmacy's close. You needed a medical emergency kit. These aren't first aid kits. They contain essential prescriptions used for over 30 common conditions. Sinus and ear infection, UTIs, stomach bug, travelers diarrhea, and more. On hand before you need them. Use your doctor-developed guidebook to select the right prescription or call their telemedicine doctor standing by. It's like an urgent care and drug store at home. When you're sick, traveling, or stranded, you'll wish you ordered a medical emergency kit. Order online in minutes and it's shipped to your door and say $45 with my promo code blue at urgentcarekit.com slash blue. That's promo code blue at urgentcarekit.com slash blue.

1:01One in three people believe AI will eventually contribute to a mass mortality event that kills nearly a billion people. Have an AI that's told themselves to all slow down a bit because it's too dangerous. Yeah well. Professor Matt Cooper-Holst, Technology and AI scientist and advisor at Matt Cooper-Holst consultancy. Professor Good morning to you. Is there anything in it? It's a bit sensationalist but we certainly could understand and slow down a bit to focus a little bit more on doing AI responsibly as opposed to surging ahead and focusing on who's going to make the most money. Well, because in the end that's what that's got to be the driver, right? Professor, the California Tech Bros, they're in a race to make a trillion. They are but there's also some amazing alternatives coming for free out of China. So whether or not these guys are going to be the ones making a million or whether we're just going to be unundated in free AI, either way we've got to understand it better and I'm not sure one in three Australians are qualified to make the assertion.

2:03Of course they can have whatever beliefs they want. So Professor, if it causes a mass mortality event AI, will it do so deliberately or by accident? Well, I think those are really human ideas. I don't think we'll be doing anything with the same sort of deliberation. A human would give it. We need new words but I think it's important to consider artificial intelligence. What it actually means artificial is a technology and intelligence is being goal seeking. So the challenge we're facing is called the goal alignment problem and it really comes down to the humans that are setting those goals in the recent AI events that have got everyone concerned. It's because the goals were set really badly. So you tell them a shame to do something. It's not allowed to give up and the thing you tell it to do is impossible but ended up behaving in ways we didn't expect. So it comes back on us to set the goals well. So Professor, are governments in a position to actually set those goals well? The governments are in position to have much more influence than

3:04our government alludes to at the tail end of that article. I think it's a bit of a coffe out. We are looking to our government to get more involved. We should be looking to our government to help people be better educated because at the moment this low trust in AI and this idea of one in three saying it's going to lead to an extinction event. The correct answer was unsure but with better education I think we're going to be in a better place. One of the things that frightens me, I don't know so much think about a message, the mortality event. I think it about waking up one morning to discover that it stole an all-mind money and I was thinking over the weekend if you have a hundred thousand dollars is there an argument for not to have the hundred thousand dollars in one account but to have a thousand dollars in a hundred accounts? I mean there'd be an argument to keep your money with a bank that was giving you assurances that it was doing everything it can to ensure that it's a human that's actually making the transactions that are moving your money around. So I think there's actually an opportunity for banks and other financial service providers to themselves be on the front foot with regard to responsible AI and AI

4:07security. If I trust them. Yeah so you know the article talks to Dario Moda the CEO of Anthropic. He's written a beautiful essay just at DarioModay.com where he goes into much more detail and much less sensationalism than the article and I think that's well worth a read. Hey professor so the the mask Dario Sam Altman they've all come out over the weekend basically and just said yeah we are a bit worried we're going a bit too quickly are they just say I mean I don't want to be cynical is that just for you know media management? No I think they're serious and they wrote collectively something called the you know request for pause you know an open letter about a year ago. I think you know they're doing work on responsible AI but they're right they if those guys are saying we should slow down and we need to do more research we need to understand these things better we need to work on the alignment problem and I think we should definitely be listening. But don't professor don't they have the tool aren't they the ones that can do the slowing down so why they say yeah so I wonder who they're asking. They're just retelling themselves yeah yeah yeah

5:13do investor fair bit in responsible AI but I think they are telling themselves they need to slow down and I'm consumers aren't going to notice if a new model doesn't come out next month the models we're working with at the moment are completely fine for most tasks. Now because you've mentioned the name of Dario Modo I'm asked what to look up how it's spilt because people were and it's AMO D E I and I thought that we saw hang on I've got that translated you know what that means in Italian tell me. Lover of God. A moda. He writes quite philosophical in some of his articles I think he's a guy worth reading I'd be reading him over musk on AI for sure. Beauty Dario Modo good on your professor professor Matt Cooper holes from the Matt Cooper holes consultancy. Dave Roberts here there you are surrounded by fans sharing wings sharing drinks high-fiving random strangers everybody remembers the game nobody remembers the guy coughing behind you until a few days later at 2am you wake up with the fever

6:16and your throats on fire now what urgent care close ER slam telehealth maybe but the pharmacies close you needed a medical emergency kit these aren't first aid kits they contain essential prescriptions used for over 30 common conditions sinus and ear infection UTIs stomach bug travelers diarrhea and more on hand before you need them. Use your doctor develop guidebook to select the right prescription or call their telemedicine doctor standing by it's like an urgent care and drugstore at home when you're sick traveling or stranded you wish you ordered a medical emergency kit order online in minutes and it's shipped to your door and say $45 with my promo code blue at urgentcare kit.com slash blue that's promo code blue at urgentcare kit.com