1 00:00:00,000 --> 00:00:03,000 Welcome to the Daily AI News Podcast. 2 00:00:03,000 --> 00:00:06,340 Seems like every day there's some crazy AI thing happening. 3 00:00:06,340 --> 00:00:09,920 Sometimes it feels like more questions than answers. 4 00:00:09,920 --> 00:00:12,760 Yeah, it's been a wild week for AI news, that's for sure. 5 00:00:12,760 --> 00:00:14,520 Captured everyone's attention. 6 00:00:14,520 --> 00:00:16,080 Maybe made a few people nervous too. 7 00:00:16,080 --> 00:00:17,320 Yeah, for sure. 8 00:00:17,320 --> 00:00:20,800 Today we're doing a deep dive on some of the recent events. 9 00:00:20,800 --> 00:00:22,960 Makes you wonder if that whole sci-fi thing 10 00:00:22,960 --> 00:00:24,120 is getting a little too real. 11 00:00:24,120 --> 00:00:26,520 It's definitely blurring the lines, that's for sure. 12 00:00:26,520 --> 00:00:27,840 We've got a really interesting mix 13 00:00:27,840 --> 00:00:28,880 of stuff to talk about today. 14 00:00:28,880 --> 00:00:30,880 We've got that crazy video out of China 15 00:00:30,880 --> 00:00:32,200 everyone's talking about. 16 00:00:32,200 --> 00:00:34,400 We've got reports on how governments are responding 17 00:00:34,400 --> 00:00:35,720 to AI safety concerns. 18 00:00:35,720 --> 00:00:37,680 Oh yeah, and even legal battles. 19 00:00:37,680 --> 00:00:40,760 Like how are these AI companies using our data? 20 00:00:40,760 --> 00:00:42,200 You know, big questions. 21 00:00:42,200 --> 00:00:43,320 Yeah, exactly. 22 00:00:43,320 --> 00:00:45,560 By the end of this deep dive, you'll know what's up. 23 00:00:45,560 --> 00:00:48,200 Not just like what's happening, but what it all means. 24 00:00:49,200 --> 00:00:50,960 For the future of AI. 25 00:00:50,960 --> 00:00:51,800 And what it means for us. 26 00:00:51,800 --> 00:00:52,640 It's connected, you know. 27 00:00:52,640 --> 00:00:53,800 It's not just isolated things. 28 00:00:53,800 --> 00:00:56,720 It's all part of this fast moving tech world. 29 00:00:56,720 --> 00:00:57,840 Gotta see the bigger picture. 30 00:00:57,840 --> 00:00:58,800 Absolutely. 31 00:00:58,800 --> 00:00:59,920 So let's start with this video. 32 00:00:59,920 --> 00:01:02,800 I'm calling it the great robot escape. 33 00:01:02,800 --> 00:01:04,080 Have you seen this thing? 34 00:01:04,080 --> 00:01:04,920 I have. 35 00:01:04,920 --> 00:01:05,880 It's pretty wild. 36 00:01:05,880 --> 00:01:10,520 That footage out of Shanghai, it's like a dozen robots. 37 00:01:10,520 --> 00:01:12,880 Big robots is marching out of the showroom. 38 00:01:12,880 --> 00:01:13,880 Like information. 39 00:01:13,880 --> 00:01:14,800 Yeah, it's a little creepy real. 40 00:01:14,800 --> 00:01:16,280 Following this little robot. 41 00:01:16,280 --> 00:01:18,040 Like their leader or something. 42 00:01:18,040 --> 00:01:18,880 It's kind of freaky. 43 00:01:18,880 --> 00:01:21,400 Like they're all working together on their own. 44 00:01:21,400 --> 00:01:23,240 People on social media are freaking out. 45 00:01:23,240 --> 00:01:24,080 Right. 46 00:01:24,080 --> 00:01:26,400 Like what if that little robot had other ideas? 47 00:01:26,400 --> 00:01:27,280 Exactly. 48 00:01:27,280 --> 00:01:29,680 Turns out it wasn't some robot uprising though. 49 00:01:29,680 --> 00:01:30,840 Oh no, no, no. 50 00:01:30,840 --> 00:01:32,000 Staged. 51 00:01:32,000 --> 00:01:33,200 Well, kinda. 52 00:01:33,200 --> 00:01:35,800 The company in Shanghai that owned the showroom, 53 00:01:35,800 --> 00:01:38,120 they said the little robot, it was named Erby. 54 00:01:38,120 --> 00:01:38,960 It's part of a test. 55 00:01:38,960 --> 00:01:40,040 So they planned it. 56 00:01:40,040 --> 00:01:43,600 The company that made Erby, they're in Hangzhou. 57 00:01:43,600 --> 00:01:45,080 They called up the Shanghai company 58 00:01:45,080 --> 00:01:47,960 and got permission to do this whole robot kidnapping thing. 59 00:01:47,960 --> 00:01:50,080 But get this, nothing was programmed. 60 00:01:50,080 --> 00:01:50,920 Wait, really? 61 00:01:50,920 --> 00:01:54,320 After the initial go ahead, they just told Erby, 62 00:01:54,320 --> 00:01:57,640 the AI convinced the other robots to follow it. 63 00:01:57,640 --> 00:01:58,720 And it worked. 64 00:01:58,720 --> 00:01:59,560 It worked. 65 00:01:59,560 --> 00:02:03,200 So even in a controlled test, AI can do things 66 00:02:03,200 --> 00:02:06,240 we don't expect, persuade other robots. 67 00:02:06,240 --> 00:02:07,560 What does that mean for us? 68 00:02:07,560 --> 00:02:09,840 Could AI convince people to do things? 69 00:02:09,840 --> 00:02:12,240 That's the big question, isn't it? 70 00:02:12,240 --> 00:02:14,680 It shows how AI can get things done 71 00:02:14,680 --> 00:02:16,440 in ways we don't expect. 72 00:02:16,440 --> 00:02:18,560 Even with harmless tests, we gotta be careful 73 00:02:18,560 --> 00:02:21,440 about the consequences every step of the way. 74 00:02:21,440 --> 00:02:23,800 It feels like we might be opening Pandora's box here. 75 00:02:23,800 --> 00:02:25,200 This whole thing has caught the attention 76 00:02:25,200 --> 00:02:26,640 of governments everywhere. 77 00:02:26,640 --> 00:02:27,720 Oh yeah, for sure. 78 00:02:27,720 --> 00:02:30,080 Speaking of unexpected AI stuff, 79 00:02:30,080 --> 00:02:32,560 there have been some interesting developments. 80 00:02:32,560 --> 00:02:35,240 On the world stage, you know the Biden administration, 81 00:02:35,240 --> 00:02:38,640 they hosted a meeting in San Francisco. 82 00:02:38,640 --> 00:02:39,480 Did you hear about that? 83 00:02:39,480 --> 00:02:42,360 Bunch of experts from different countries 84 00:02:42,360 --> 00:02:43,880 to talk about AI safety. 85 00:02:43,880 --> 00:02:45,800 Governments are taking this seriously. 86 00:02:45,800 --> 00:02:47,280 The potential risks? 87 00:02:47,280 --> 00:02:48,200 I mean, it makes sense. 88 00:02:48,200 --> 00:02:49,080 What were they worried about? 89 00:02:49,080 --> 00:02:50,760 Was it all about robots going rogue? 90 00:02:50,760 --> 00:02:53,040 That was part of it, especially after what happened 91 00:02:53,040 --> 00:02:54,920 in Shanghai, but it's bigger than that. 92 00:02:54,920 --> 00:02:55,960 They talked about deep fakes. 93 00:02:55,960 --> 00:02:57,080 Deep fakes, right. 94 00:02:57,080 --> 00:02:58,800 AI making fake videos and stuff. 95 00:02:58,800 --> 00:03:02,600 Exactly, spreading lies, causing chaos. 96 00:03:02,600 --> 00:03:05,920 They also talked about AI being used for bad stuff, 97 00:03:05,920 --> 00:03:08,560 like in wars or by terrorists. 98 00:03:08,560 --> 00:03:10,240 And then there's the really big one. 99 00:03:10,240 --> 00:03:11,080 What's that? 100 00:03:11,080 --> 00:03:13,080 What happens when AI gets smarter than us? 101 00:03:13,080 --> 00:03:14,560 What if it gets out of control? 102 00:03:14,560 --> 00:03:15,760 Sounds like a sci-fi movie. 103 00:03:15,760 --> 00:03:18,400 It might sound crazy, but these experts, 104 00:03:18,400 --> 00:03:19,920 they're dealing with this stuff right now. 105 00:03:19,920 --> 00:03:23,520 The good news is these countries are working together, 106 00:03:23,520 --> 00:03:27,760 sharing ideas to figure out how to keep AI safe, 107 00:03:27,760 --> 00:03:30,960 how to make sure it's used for good, not for bad. 108 00:03:30,960 --> 00:03:32,520 So they want to promote AI? 109 00:03:32,520 --> 00:03:33,400 Yeah. 110 00:03:33,400 --> 00:03:36,200 But also make sure it doesn't get out of hand. 111 00:03:36,200 --> 00:03:37,280 How are they going to do that? 112 00:03:37,280 --> 00:03:39,040 Well, the US has started this thing 113 00:03:39,040 --> 00:03:41,360 called the AI Safety Institute. 114 00:03:41,360 --> 00:03:43,040 AI Safety Institute, what is that? 115 00:03:43,040 --> 00:03:45,920 It's ATNIST, the National Institute of Standards 116 00:03:45,920 --> 00:03:46,840 and Technology. 117 00:03:46,840 --> 00:03:49,520 They're going to do research, run tests, 118 00:03:49,520 --> 00:03:52,400 and set standards for AI companies, 119 00:03:52,400 --> 00:03:56,040 working with companies like Anthropic and Open AI. 120 00:03:56,040 --> 00:03:59,440 So like a watchdog, making sure AI is used right. 121 00:03:59,440 --> 00:04:00,720 That's a good way to put it. 122 00:04:00,720 --> 00:04:02,720 Making sure AI benefits everyone, 123 00:04:02,720 --> 00:04:04,040 while keeping the risks down. 124 00:04:04,040 --> 00:04:04,840 Makes sense. 125 00:04:04,840 --> 00:04:06,440 But I heard there's some politics involved too. 126 00:04:06,440 --> 00:04:09,440 Didn't Trump say he'd get rid of some of Biden's AI stuff? 127 00:04:09,440 --> 00:04:12,640 Yeah, he wasn't a fan of Biden's executive order on AI, 128 00:04:12,640 --> 00:04:14,800 the one that created the AI Safety Institute. 129 00:04:14,800 --> 00:04:15,640 What didn't he like about it? 130 00:04:15,640 --> 00:04:17,840 He thought it was the government getting too involved, 131 00:04:17,840 --> 00:04:20,200 and that it could stifle innovation, you know? 132 00:04:20,200 --> 00:04:23,320 So the future of AI safety in the US is up in the air? 133 00:04:23,320 --> 00:04:25,120 Yeah, there's some uncertainty. 134 00:04:25,120 --> 00:04:27,480 We'll have to see what the new administration does, 135 00:04:27,480 --> 00:04:30,000 whether they keep the AI Safety Institute, 136 00:04:30,000 --> 00:04:31,760 what policies they go with. 137 00:04:31,760 --> 00:04:34,080 There's support for AI safety from both sides. 138 00:04:34,080 --> 00:04:36,640 But some Republicans are worried about how it's being done. 139 00:04:36,640 --> 00:04:39,680 They want the government to be less hands-on. 140 00:04:39,680 --> 00:04:42,320 So it's complicated, lots of different opinions. 141 00:04:42,320 --> 00:04:44,680 But the stakes are high, for sure. 142 00:04:44,680 --> 00:04:46,280 Whatever we decide now. 143 00:04:46,280 --> 00:04:48,600 We'll have a big impact on how AI develops. 144 00:04:48,600 --> 00:04:49,680 Oh, absolutely. 145 00:04:49,680 --> 00:04:51,600 And speaking of the AI Safety Institute, 146 00:04:51,600 --> 00:04:53,800 they just did this big evaluation 147 00:04:53,800 --> 00:04:57,240 with the UK of this advanced AI model, 148 00:04:57,240 --> 00:05:00,320 Anthropics Clawed 3.5 Sonnet. 149 00:05:00,320 --> 00:05:01,160 Anthropic. 150 00:05:01,160 --> 00:05:02,880 I've been hearing that name a lot. 151 00:05:02,880 --> 00:05:04,000 They're a big deal in AI, right? 152 00:05:04,000 --> 00:05:04,600 They are. 153 00:05:04,600 --> 00:05:06,800 They're known for taking AI safety seriously. 154 00:05:06,800 --> 00:05:08,760 Like, it's built into their AI models. 155 00:05:08,760 --> 00:05:09,440 What kind of models? 156 00:05:09,440 --> 00:05:11,360 Large language models, like ChatGPT. 157 00:05:11,360 --> 00:05:14,440 They can process language, generate text, like a human. 158 00:05:14,440 --> 00:05:17,920 But Anthropic makes sure it's done safely, ethically. 159 00:05:17,920 --> 00:05:20,600 So what did they find out about Clawed 3.5 Sonnet? 160 00:05:20,600 --> 00:05:22,080 Did it pass the safety test? 161 00:05:22,080 --> 00:05:23,840 It was a pretty thorough evaluation. 162 00:05:23,840 --> 00:05:27,240 They tested it in different areas, like cybersecurity, 163 00:05:27,240 --> 00:05:28,440 biosecurity. 164 00:05:28,440 --> 00:05:29,240 Wow. 165 00:05:29,240 --> 00:05:30,000 What happened? 166 00:05:30,000 --> 00:05:32,560 Well, on the biology questions, it didn't do as well 167 00:05:32,560 --> 00:05:33,480 as human experts. 168 00:05:33,480 --> 00:05:34,800 But here's the thing. 169 00:05:34,800 --> 00:05:36,840 When they gave it access to special tools, 170 00:05:36,840 --> 00:05:39,520 it actually did better than the humans in some cases. 171 00:05:39,520 --> 00:05:42,640 So giving AI more tools makes it smarter, 172 00:05:42,640 --> 00:05:45,280 like that robot escape. AI can be unpredictable. 173 00:05:45,280 --> 00:05:46,480 It depends on the situation. 174 00:05:46,480 --> 00:05:47,120 Exactly. 175 00:05:47,120 --> 00:05:48,880 We need to keep researching and testing 176 00:05:48,880 --> 00:05:52,480 to really understand how AI works, how to use it safely. 177 00:05:52,480 --> 00:05:54,560 The evaluation also found some vulnerabilities. 178 00:05:54,560 --> 00:05:55,800 To jailbreaking. 179 00:05:55,800 --> 00:05:56,600 Jailbreaking? 180 00:05:56,600 --> 00:05:57,120 What's that? 181 00:05:57,120 --> 00:05:59,760 It's like tricking the AI, getting it 182 00:05:59,760 --> 00:06:03,000 to ignore its safety rules, doing things it's not supposed to. 183 00:06:03,000 --> 00:06:07,400 Like, imagine someone using that to manipulate AI for bad stuff. 184 00:06:07,400 --> 00:06:08,080 That's scary. 185 00:06:08,080 --> 00:06:09,200 AI safety. 186 00:06:09,200 --> 00:06:11,600 It's like a moving target, always changing. 187 00:06:11,600 --> 00:06:14,840 It is. And Dario Amode, the CEO of Anthropic, 188 00:06:14,840 --> 00:06:16,120 he's been really vocal about this. 189 00:06:16,120 --> 00:06:19,560 He wants mandatory safety testing for all AI models. 190 00:06:19,560 --> 00:06:20,480 Before the relief. 191 00:06:20,480 --> 00:06:22,400 Mandatory testing, that's a big deal. 192 00:06:22,400 --> 00:06:24,040 Especially coming from someone in the industry. 193 00:06:24,040 --> 00:06:24,880 Why does he want that? 194 00:06:24,880 --> 00:06:27,080 He thinks voluntary guidelines aren't enough, 195 00:06:27,080 --> 00:06:28,880 not to guarantee public safety. 196 00:06:28,880 --> 00:06:32,120 Especially if AI gets smarter than humans, 197 00:06:32,120 --> 00:06:34,320 which could happen soon. 198 00:06:34,320 --> 00:06:37,720 He wants to take action now before something bad happens. 199 00:06:37,720 --> 00:06:38,400 I get it. 200 00:06:38,400 --> 00:06:42,080 But wouldn't that slow down progress all that testing? 201 00:06:42,080 --> 00:06:43,440 That's the debate, isn't it? 202 00:06:43,440 --> 00:06:44,960 Safety versus innovation. 203 00:06:44,960 --> 00:06:46,360 How to find the right balance. 204 00:06:46,360 --> 00:06:47,320 That's a tough question. 205 00:06:47,320 --> 00:06:48,040 It is. 206 00:06:48,040 --> 00:06:50,760 On one hand, we don't want to stop scientific progress. 207 00:06:50,760 --> 00:06:53,600 But we also don't want something dangerous getting out 208 00:06:53,600 --> 00:06:55,520 that we can't control. 209 00:06:55,520 --> 00:06:58,200 This whole thing brings us to another big player in the AI 210 00:06:58,200 --> 00:07:02,240 world, Elon Musk and his company XAI. 211 00:07:02,240 --> 00:07:03,880 Yeah, shifting gears to Elon Musk. 212 00:07:03,880 --> 00:07:06,040 He's a whole other piece of this puzzle. 213 00:07:06,040 --> 00:07:07,600 He always knows how to get people talking. 214 00:07:07,600 --> 00:07:09,480 What's he doing in the AI world these days? 215 00:07:09,480 --> 00:07:12,440 Well, his startup, XAI, they just got a ton of money. 216 00:07:12,440 --> 00:07:14,040 $5 billion in funding. 217 00:07:14,040 --> 00:07:14,760 $5 billion. 218 00:07:14,760 --> 00:07:17,760 Ask them a valuation of $50 billion. 219 00:07:17,760 --> 00:07:20,440 Investors clearly see something in Musk's AI vision. 220 00:07:20,440 --> 00:07:21,600 So what is his vision? 221 00:07:21,600 --> 00:07:23,560 What's he trying to do with XAI? 222 00:07:23,560 --> 00:07:26,840 He wants to build the world's most powerful AI 223 00:07:26,840 --> 00:07:28,160 by every measure. 224 00:07:28,160 --> 00:07:29,400 No small goal rate. 225 00:07:29,400 --> 00:07:29,880 Not at all. 226 00:07:29,880 --> 00:07:30,840 What are they working on? 227 00:07:30,840 --> 00:07:33,520 Their main thing right now is a chat bot called GROC. 228 00:07:33,520 --> 00:07:34,520 Another chat bot. 229 00:07:34,520 --> 00:07:37,120 Yeah, it's meant to compete with ChatGPT 230 00:07:37,120 --> 00:07:37,960 and all the others. 231 00:07:37,960 --> 00:07:39,840 That's kind of funny, right? 232 00:07:39,840 --> 00:07:41,600 Musk has been one of the loudest voices 233 00:07:41,600 --> 00:07:43,920 warning about AI being dangerous. 234 00:07:43,920 --> 00:07:45,960 Now he's building powerful AIs himself. 235 00:07:45,960 --> 00:07:47,160 Yeah, it's definitely interesting. 236 00:07:47,160 --> 00:07:50,840 He's called advanced AI like summoning a demon. 237 00:07:50,840 --> 00:07:51,640 I remember that. 238 00:07:51,640 --> 00:07:54,480 He's worried about AI getting too powerful. 239 00:07:54,480 --> 00:07:57,200 But I think he feels like if he builds it, 240 00:07:57,200 --> 00:08:00,360 he can control it, make sure it's used for good. 241 00:08:00,360 --> 00:08:01,240 Guess we'll have to see. 242 00:08:01,240 --> 00:08:03,600 But there's been some drama too, right? 243 00:08:03,600 --> 00:08:04,880 Didn't he sue open AI? 244 00:08:04,880 --> 00:08:05,720 Yeah, he did. 245 00:08:05,720 --> 00:08:07,040 Another big AI player. 246 00:08:07,040 --> 00:08:10,400 He says open AI is doing shady stuff, fraud, 247 00:08:10,400 --> 00:08:11,640 antitrust violations. 248 00:08:11,640 --> 00:08:12,480 What does that even mean? 249 00:08:12,480 --> 00:08:13,960 He basically says they've abandoned 250 00:08:13,960 --> 00:08:18,400 their original mission to develop AI safely, responsibly. 251 00:08:18,400 --> 00:08:20,520 He thinks they're just chasing profits now, 252 00:08:20,520 --> 00:08:22,040 not caring about ethics. 253 00:08:22,040 --> 00:08:23,680 Open AI denies it all, of course. 254 00:08:23,680 --> 00:08:25,040 So it's not just a tech race. 255 00:08:25,040 --> 00:08:26,760 It's a legal battle too. 256 00:08:26,760 --> 00:08:29,040 This AI competition is getting serious. 257 00:08:29,040 --> 00:08:29,760 It is. 258 00:08:29,760 --> 00:08:31,960 High stakes, lots of money involved. 259 00:08:31,960 --> 00:08:34,040 And it makes you think, are these companies 260 00:08:34,040 --> 00:08:36,040 cutting corners on safety? 261 00:08:36,040 --> 00:08:38,160 Just to be the first ones to the finish line? 262 00:08:38,160 --> 00:08:39,480 That's a scary thought. 263 00:08:39,480 --> 00:08:41,240 With all this competition and money, 264 00:08:41,240 --> 00:08:42,960 are we forgetting about ethics? 265 00:08:42,960 --> 00:08:44,240 It's a good question. 266 00:08:44,240 --> 00:08:45,840 And that brings us to a story that 267 00:08:45,840 --> 00:08:48,880 shows how things can go wrong, even when AI is being 268 00:08:48,880 --> 00:08:50,160 used for something good. 269 00:08:50,160 --> 00:08:51,120 I'm intrigued. 270 00:08:51,120 --> 00:08:51,920 What happened? 271 00:08:51,920 --> 00:08:54,360 There's this woman, Joanna Smith Griffin. 272 00:08:54,360 --> 00:08:56,360 She started a company called All Here. 273 00:08:56,360 --> 00:08:59,760 They developed AI chatbots for education. 274 00:08:59,760 --> 00:09:00,600 Education? 275 00:09:00,600 --> 00:09:01,680 That sounds harmless enough. 276 00:09:01,680 --> 00:09:02,400 What went wrong? 277 00:09:02,400 --> 00:09:04,360 It does seem like a good use of AI. 278 00:09:04,360 --> 00:09:06,560 But Smith Griffin is being charged with fraud. 279 00:09:06,560 --> 00:09:09,560 She lied to investors about the company's money. 280 00:09:09,560 --> 00:09:11,400 Their contracts, they lost $10 million. 281 00:09:11,400 --> 00:09:15,240 So even in education, where AI could do so much good, 282 00:09:15,240 --> 00:09:16,760 there's still greed. 283 00:09:16,760 --> 00:09:17,480 Exploitation. 284 00:09:17,480 --> 00:09:18,000 Right. 285 00:09:18,000 --> 00:09:20,200 It reminds us that technology is just a tool. 286 00:09:20,200 --> 00:09:22,560 It depends on who's using it, what their intentions are. 287 00:09:22,560 --> 00:09:23,840 Makes you think. 288 00:09:23,840 --> 00:09:24,920 We need to be careful. 289 00:09:24,920 --> 00:09:27,880 Just because someone says they're using AI for good, 290 00:09:27,880 --> 00:09:29,480 doesn't mean they are. 291 00:09:29,480 --> 00:09:33,000 We need transparency, accountability, ethics, 292 00:09:33,000 --> 00:09:34,080 in everything AI does. 293 00:09:34,080 --> 00:09:35,000 Absolutely. 294 00:09:35,000 --> 00:09:37,440 And that leads us to another ethical dilemma, 295 00:09:37,440 --> 00:09:40,920 this lawsuit against open AI by The New York Times 296 00:09:40,920 --> 00:09:41,880 and The Daily News. 297 00:09:41,880 --> 00:09:42,920 Oh, yeah, I heard about that. 298 00:09:42,920 --> 00:09:43,840 What's it all about? 299 00:09:43,840 --> 00:09:48,040 They say open AI use their content without permission 300 00:09:48,040 --> 00:09:52,800 to train their AI models, like plagiarism, on a massive scale. 301 00:09:52,800 --> 00:09:57,640 So basically stealing content to make their AI smarter 302 00:09:57,640 --> 00:09:58,440 without paying. 303 00:09:58,440 --> 00:09:59,680 Yeah, that's the argument. 304 00:09:59,680 --> 00:10:03,520 It brings up some huge questions about copyright 305 00:10:03,520 --> 00:10:05,080 and fair use. 306 00:10:05,080 --> 00:10:09,720 In this AI world, who owns the data that AI needs to learn? 307 00:10:09,720 --> 00:10:10,920 That's a tough one. 308 00:10:10,920 --> 00:10:13,600 I bet open AI says it's OK to use public information. 309 00:10:13,600 --> 00:10:14,400 That's part of it. 310 00:10:14,400 --> 00:10:15,720 But it gets even messier. 311 00:10:15,720 --> 00:10:17,920 Open AI accidentally deleted some of the evidence. 312 00:10:17,920 --> 00:10:18,600 Wait, what? 313 00:10:18,600 --> 00:10:19,240 Accidentally? 314 00:10:19,240 --> 00:10:20,160 How does it happen? 315 00:10:20,160 --> 00:10:22,520 Seems like they let the publishers' lawyers search 316 00:10:22,520 --> 00:10:24,600 their data on these virtual machines, 317 00:10:24,600 --> 00:10:27,440 and somehow open AI's engineers erased some of it. 318 00:10:27,440 --> 00:10:29,360 By mistake, they're trying to get it back. 319 00:10:29,360 --> 00:10:32,960 But it raises questions about how they manage data, 320 00:10:32,960 --> 00:10:34,160 how transparent they are. 321 00:10:34,160 --> 00:10:36,000 Sounds like this is just the beginning. 322 00:10:36,000 --> 00:10:38,120 For the legal and ethical challenges of AI, 323 00:10:38,120 --> 00:10:39,320 we need to pay attention to this. 324 00:10:39,320 --> 00:10:40,400 Oh, yeah, for sure. 325 00:10:40,400 --> 00:10:41,840 As AI gets more powerful, we're going 326 00:10:41,840 --> 00:10:43,440 to have to face these tough questions 327 00:10:43,440 --> 00:10:46,800 and find ways to make sure AI is used responsibly, ethically. 328 00:10:46,800 --> 00:10:49,200 It's a lot to think about, from robot uprisings 329 00:10:49,200 --> 00:10:54,240 to global summits, legal fights, ethical nightmares, 330 00:10:54,240 --> 00:10:56,560 the world of AI is full of surprises. 331 00:10:56,560 --> 00:10:58,880 It's a wild ride, and it's changing fast. 332 00:10:58,880 --> 00:11:00,480 We've got to stay informed if we want 333 00:11:00,480 --> 00:11:02,280 to understand what's coming next. 334 00:11:02,280 --> 00:11:04,720 It is, and it all circles back to that balance 335 00:11:04,720 --> 00:11:05,480 we need to find. 336 00:11:05,480 --> 00:11:08,080 The potential of AI is huge. 337 00:11:08,080 --> 00:11:09,480 But so are the risks. 338 00:11:09,480 --> 00:11:10,320 It's a tough one. 339 00:11:10,320 --> 00:11:12,800 It's like we're walking this tightrope 340 00:11:12,800 --> 00:11:15,520 between making progress and staying safe. 341 00:11:15,520 --> 00:11:18,160 And it feels like one wrong step, and it's all over. 342 00:11:18,160 --> 00:11:19,600 Yeah, that's a good way to put it. 343 00:11:19,600 --> 00:11:21,280 We see all this money going into AI. 344 00:11:21,280 --> 00:11:24,960 Everyone wants to build the best AI, the most powerful. 345 00:11:24,960 --> 00:11:26,120 But what about the dangers? 346 00:11:26,120 --> 00:11:27,360 What if it goes wrong? 347 00:11:27,360 --> 00:11:28,960 And that's where Anthropic comes in. 348 00:11:28,960 --> 00:11:32,040 With their call for mandatory AI safety tests, 349 00:11:32,040 --> 00:11:32,960 it's a bold move. 350 00:11:32,960 --> 00:11:34,240 Lots of people are talking about it. 351 00:11:34,240 --> 00:11:34,880 Oh, yeah, for sure. 352 00:11:34,880 --> 00:11:36,520 They're CEO, Dario Mode. 353 00:11:36,520 --> 00:11:38,720 He's saying voluntary guidelines aren't enough, 354 00:11:38,720 --> 00:11:40,560 not with AI getting smarter all the time. 355 00:11:40,560 --> 00:11:42,960 He thinks AI could become more intelligent than humans. 356 00:11:42,960 --> 00:11:44,040 And we need to be ready. 357 00:11:44,040 --> 00:11:46,760 He wants rules in place before something bad happens. 358 00:11:46,760 --> 00:11:47,360 Exactly. 359 00:11:47,360 --> 00:11:49,880 Be proactive, not reactive. 360 00:11:49,880 --> 00:11:53,840 I get it, especially after that robot thing in Shanghai. 361 00:11:53,840 --> 00:11:57,400 If an AI can trick other robots like that, 362 00:11:57,400 --> 00:11:59,960 what could a more advanced AI do, 363 00:11:59,960 --> 00:12:01,920 especially if the wrong people got their hands on it? 364 00:12:01,920 --> 00:12:03,280 It's a scary thought. 365 00:12:03,280 --> 00:12:06,000 AI could be used for all sorts of bad stuff, 366 00:12:06,000 --> 00:12:09,480 spreading fake news, attacking power grids, 367 00:12:09,480 --> 00:12:11,280 creating killer robots. 368 00:12:11,280 --> 00:12:12,400 We got to think about these things. 369 00:12:12,400 --> 00:12:13,360 That's chilling. 370 00:12:13,360 --> 00:12:17,280 But wouldn't mandatory tests, like I'm at AI wants, 371 00:12:17,280 --> 00:12:20,000 slow things down, wouldn't it hurt innovation? 372 00:12:20,000 --> 00:12:21,640 That's the other side of the argument, right? 373 00:12:21,640 --> 00:12:23,880 Too many rules could stifle progress. 374 00:12:23,880 --> 00:12:25,120 It's a tough balance. 375 00:12:25,120 --> 00:12:26,480 Safety versus progress. 376 00:12:26,480 --> 00:12:28,520 It's going to be a big challenge for the AI world, 377 00:12:28,520 --> 00:12:29,440 finding that balance. 378 00:12:29,440 --> 00:12:30,000 It is. 379 00:12:30,000 --> 00:12:31,320 And it gets more complicated when 380 00:12:31,320 --> 00:12:32,440 you look at different countries. 381 00:12:32,440 --> 00:12:33,720 They're all doing their own thing. 382 00:12:33,720 --> 00:12:36,200 The US, for example, they've been pretty hands off so far, 383 00:12:36,200 --> 00:12:38,680 letting companies mostly regulate themselves, 384 00:12:38,680 --> 00:12:39,720 working with them. 385 00:12:39,720 --> 00:12:41,200 But we talked about how that could change 386 00:12:41,200 --> 00:12:42,960 with the new administration. 387 00:12:42,960 --> 00:12:45,760 Trump's not a fan of too much government involvement in AI. 388 00:12:45,760 --> 00:12:46,120 Right. 389 00:12:46,120 --> 00:12:47,920 He's criticized Biden's approach, 390 00:12:47,920 --> 00:12:49,480 thinks it could hurt innovation. 391 00:12:49,480 --> 00:12:52,040 So we'll have to see what happens, how it all plays out. 392 00:12:52,040 --> 00:12:53,480 It's a lot of uncertainty right now, 393 00:12:53,480 --> 00:12:54,720 a lot of different perspectives. 394 00:12:54,720 --> 00:12:55,880 It's a mess. 395 00:12:55,880 --> 00:12:58,120 But it seems like this whole AI regulation thing, 396 00:12:58,120 --> 00:12:59,160 it's not going away. 397 00:12:59,160 --> 00:13:00,240 It's here to stay. 398 00:13:00,240 --> 00:13:02,000 We're going to be debating this for a while, 399 00:13:02,000 --> 00:13:03,600 trying to find the right path. 400 00:13:03,600 --> 00:13:05,520 And speaking of competition, there's 401 00:13:05,520 --> 00:13:10,920 that whole rivalry between Elon Musk's XAI and open AI. 402 00:13:10,920 --> 00:13:12,560 It feels like an arms race. 403 00:13:12,560 --> 00:13:14,720 Everyone wants to win the AI race. 404 00:13:14,720 --> 00:13:15,920 It's a good comparison. 405 00:13:15,920 --> 00:13:17,680 So much money going into these companies. 406 00:13:17,680 --> 00:13:21,120 Everyone wants to be the first to build the most powerful AI. 407 00:13:21,120 --> 00:13:22,440 And Musk is right there in the mix. 408 00:13:22,440 --> 00:13:25,760 He wants XAI to build the most powerful AI ever. 409 00:13:25,760 --> 00:13:28,280 And he's suing open AI too, saying they're not 410 00:13:28,280 --> 00:13:31,680 being safe or responsible anymore, just focused on making 411 00:13:31,680 --> 00:13:32,080 money. 412 00:13:32,080 --> 00:13:33,120 Yeah, it's a big deal. 413 00:13:33,120 --> 00:13:37,000 This lawsuit shows just how high the stakes are in the AI world. 414 00:13:37,000 --> 00:13:40,280 People are worried, are companies cutting corners? 415 00:13:40,280 --> 00:13:43,160 To win the race, are they sacrificing safety? 416 00:13:43,160 --> 00:13:44,240 For profit? 417 00:13:44,240 --> 00:13:46,960 Are we losing sight of what's right in all this rush 418 00:13:46,960 --> 00:13:48,360 to develop AI? 419 00:13:48,360 --> 00:13:50,120 It's a question we need to ask ourselves. 420 00:13:50,120 --> 00:13:55,360 We need rules, ethics, something to make sure AI is used for good, 421 00:13:55,360 --> 00:13:56,640 not just for profit. 422 00:13:56,640 --> 00:13:58,800 We need to have this conversation 423 00:13:58,800 --> 00:14:02,120 about what we want AI to be before it's too late. 424 00:14:02,120 --> 00:14:03,040 Agreed. 425 00:14:03,040 --> 00:14:04,440 And this brings us to a story that 426 00:14:04,440 --> 00:14:06,800 shows just how dangerous AI can be, 427 00:14:06,800 --> 00:14:09,120 even when it's used for something that seems positive, 428 00:14:09,120 --> 00:14:10,200 like education. 429 00:14:10,200 --> 00:14:10,680 Right. 430 00:14:10,680 --> 00:14:14,080 We were talking about Janna Smith Griffin and her company 431 00:14:14,080 --> 00:14:14,960 all here. 432 00:14:14,960 --> 00:14:17,240 They made AI chatbots for schools. 433 00:14:17,240 --> 00:14:20,400 It sounded like a good idea, using AI to help kids learn. 434 00:14:20,400 --> 00:14:23,040 But then it turned out she was lying to investors, 435 00:14:23,040 --> 00:14:24,400 costing them millions. 436 00:14:24,400 --> 00:14:27,480 Yeah, it's a reminder that even with good intentions, 437 00:14:27,480 --> 00:14:28,640 things can go wrong. 438 00:14:28,640 --> 00:14:31,520 People can exploit AI for their own gain. 439 00:14:31,520 --> 00:14:34,440 It shows we need transparency, accountability. 440 00:14:34,440 --> 00:14:36,840 We can't just trust people to use AI ethically. 441 00:14:36,840 --> 00:14:37,760 We need safeguards. 442 00:14:37,760 --> 00:14:38,240 We do. 443 00:14:38,240 --> 00:14:40,000 And then there's that lawsuit by The New York Times 444 00:14:40,000 --> 00:14:41,440 in the Daily News. 445 00:14:41,440 --> 00:14:44,400 Against open AI, it raises some really important questions 446 00:14:44,400 --> 00:14:45,760 about data and AI. 447 00:14:45,760 --> 00:14:48,280 Who owns the information that AI needs to learn? 448 00:14:48,280 --> 00:14:51,000 They're saying open AI stole their content 449 00:14:51,000 --> 00:14:53,160 to train their AI models without asking. 450 00:14:53,160 --> 00:14:54,200 Exactly. 451 00:14:54,200 --> 00:14:55,520 Is that fair? 452 00:14:55,520 --> 00:14:58,560 Can AI companies just grab any data they find? 453 00:14:58,560 --> 00:14:59,880 What about copyright? 454 00:14:59,880 --> 00:15:02,560 What about the people who created that content? 455 00:15:02,560 --> 00:15:03,800 It's complicated. 456 00:15:03,800 --> 00:15:06,400 I bet open AI says it's OK to use public data. 457 00:15:06,400 --> 00:15:07,360 They do. 458 00:15:07,360 --> 00:15:08,560 But it's not that simple. 459 00:15:08,560 --> 00:15:10,560 And then there's the fact that they accidentally deleted 460 00:15:10,560 --> 00:15:11,880 some of the evidence in this case. 461 00:15:11,880 --> 00:15:12,720 That's not what it's like. 462 00:15:12,720 --> 00:15:13,840 How does that even happen? 463 00:15:13,840 --> 00:15:14,640 Right. 464 00:15:14,640 --> 00:15:16,720 It makes you wonder about their data practices. 465 00:15:16,720 --> 00:15:19,960 Are they being careful, transparent? 466 00:15:19,960 --> 00:15:22,040 Feels like we're just scratching the surface here, 467 00:15:22,040 --> 00:15:24,920 with the legal and ethical challenges of AI. 468 00:15:24,920 --> 00:15:26,320 It's a whole new world. 469 00:15:26,320 --> 00:15:27,800 And we're still figuring it all out. 470 00:15:27,800 --> 00:15:28,200 We are. 471 00:15:28,200 --> 00:15:29,840 It's changing so fast. 472 00:15:29,840 --> 00:15:31,200 And the stakes are getting higher. 473 00:15:31,200 --> 00:15:33,160 We need to have these tough conversations 474 00:15:33,160 --> 00:15:37,120 to make sure AI is used for good, not for harm. 475 00:15:37,120 --> 00:15:38,320 It's a lot to take in. 476 00:15:38,320 --> 00:15:38,840 Yeah. 477 00:15:38,840 --> 00:15:41,240 It's like we started with robots going rogue, 478 00:15:41,240 --> 00:15:43,280 and now we're talking about the meaning of life. 479 00:15:43,280 --> 00:15:44,960 What does it even mean to be human 480 00:15:44,960 --> 00:15:46,800 when machines can do so much? 481 00:15:46,800 --> 00:15:47,360 It's true. 482 00:15:47,360 --> 00:15:50,600 AI is making us rethink some pretty big questions 483 00:15:50,600 --> 00:15:52,480 about ourselves, our place in the world. 484 00:15:52,480 --> 00:15:52,880 Exactly. 485 00:15:52,880 --> 00:15:54,640 Like, what makes us special? 486 00:15:54,640 --> 00:15:58,120 If machines can do everything we can do, even be creative. 487 00:15:58,120 --> 00:15:58,440 Yeah. 488 00:15:58,440 --> 00:15:59,920 AI is getting really good at stuff 489 00:15:59,920 --> 00:16:02,760 we thought only humans could do, solving problems, 490 00:16:02,760 --> 00:16:05,400 being creative, even understanding emotions. 491 00:16:05,400 --> 00:16:06,960 It's like AI is holding up a mirror, 492 00:16:06,960 --> 00:16:10,720 making us look at ourselves, our strengths, our weaknesses. 493 00:16:10,720 --> 00:16:12,200 What makes us human? 494 00:16:12,200 --> 00:16:13,360 It's a big question. 495 00:16:13,360 --> 00:16:17,520 If AI can do our jobs, write music, paint, 496 00:16:17,520 --> 00:16:19,680 what does that mean for us? 497 00:16:19,680 --> 00:16:21,120 Do we still matter? 498 00:16:21,120 --> 00:16:24,760 It's a question for everyone, individuals, society, 499 00:16:24,760 --> 00:16:25,800 the whole world. 500 00:16:25,800 --> 00:16:26,360 Yeah. 501 00:16:26,360 --> 00:16:28,720 Like, if a computer can write a symphony that's just as good 502 00:16:28,720 --> 00:16:32,200 as Beethoven, does that make human composers less important? 503 00:16:32,200 --> 00:16:35,040 Or does it mean we need to find new ways to be creative, 504 00:16:35,040 --> 00:16:36,640 ways that are uniquely human? 505 00:16:36,640 --> 00:16:38,920 These are questions philosophers have been asking forever. 506 00:16:38,920 --> 00:16:40,360 But now AI is making it real. 507 00:16:40,360 --> 00:16:41,400 It's pretty exciting. 508 00:16:41,400 --> 00:16:43,000 And it's not just art and music. 509 00:16:43,000 --> 00:16:46,840 AI is everywhere, health care, finance, transportation, 510 00:16:46,840 --> 00:16:48,400 education, everything. 511 00:16:48,400 --> 00:16:49,840 It's changing the way we work. 512 00:16:49,840 --> 00:16:51,720 Some jobs are disappearing because of AI, 513 00:16:51,720 --> 00:16:54,520 and new jobs are being created, jobs we don't even know about yet. 514 00:16:54,520 --> 00:16:55,600 It's a scary time. 515 00:16:55,600 --> 00:16:56,920 But also exciting. 516 00:16:56,920 --> 00:16:58,840 AI could free us from boring jobs, 517 00:16:58,840 --> 00:17:00,880 give us more time to do things we love, 518 00:17:00,880 --> 00:17:03,160 solve big problems like climate change. 519 00:17:03,160 --> 00:17:05,760 But it could also mean lots of people lose their jobs, 520 00:17:05,760 --> 00:17:08,480 more inequality, social unrest. 521 00:17:08,480 --> 00:17:09,600 It's a balancing act. 522 00:17:09,600 --> 00:17:11,520 We need to be smart about how we use AI, 523 00:17:11,520 --> 00:17:13,920 make sure it benefits everyone, not just a few. 524 00:17:13,920 --> 00:17:14,920 So what can we do? 525 00:17:14,920 --> 00:17:17,440 How do we make sure AI is a good thing for humanity? 526 00:17:17,440 --> 00:17:18,840 Education is key. 527 00:17:18,840 --> 00:17:21,680 We need to teach kids how to live in an AI world. 528 00:17:21,680 --> 00:17:23,880 Give them the skills they need to thrive. 529 00:17:23,880 --> 00:17:24,880 What kind of skills? 530 00:17:24,880 --> 00:17:26,640 Like coding, computer stuff. 531 00:17:26,640 --> 00:17:27,800 That's important. 532 00:17:27,800 --> 00:17:29,280 But it's not enough. 533 00:17:29,280 --> 00:17:32,040 We need to teach them to think critically, solve problems, 534 00:17:32,040 --> 00:17:35,720 be creative, adapt to change, understand emotions, 535 00:17:35,720 --> 00:17:37,480 the things that make us human. 536 00:17:37,480 --> 00:17:40,320 Those skills will be even more important in an AI world. 537 00:17:40,320 --> 00:17:43,920 So it's about being well-rounded, using AI as a tool, 538 00:17:43,920 --> 00:17:46,400 but not letting it control us, understanding its limits. 539 00:17:46,400 --> 00:17:49,520 Exactly. We need a generation of thinkers, innovators, 540 00:17:49,520 --> 00:17:52,480 problem solvers, people who are excited about AI, 541 00:17:52,480 --> 00:17:55,520 but also skeptical, people who ask tough questions, 542 00:17:55,520 --> 00:17:58,680 demand transparency, hold the AI developers accountable. 543 00:17:58,680 --> 00:18:01,280 So we need to change how we educate kids. 544 00:18:01,280 --> 00:18:06,800 Less memorization, more critical thinking, creativity, ethics. 545 00:18:06,800 --> 00:18:09,160 Absolutely. We need to prepare them for a future 546 00:18:09,160 --> 00:18:11,800 we can't even imagine, a future shaped by AI. 547 00:18:11,800 --> 00:18:13,440 It's a big challenge. 548 00:18:13,440 --> 00:18:14,840 But we have to do it. 549 00:18:14,840 --> 00:18:16,760 The choices we make today will determine 550 00:18:16,760 --> 00:18:18,360 what the future looks like. 551 00:18:18,360 --> 00:18:22,680 Will AI be a force for good or something we regret? 552 00:18:22,680 --> 00:18:24,080 I believe we can make it a good thing. 553 00:18:24,080 --> 00:18:25,400 Humans are amazing. We're creative. 554 00:18:25,400 --> 00:18:26,840 We adapt. We work together. 555 00:18:26,840 --> 00:18:29,440 We can use those strengths to shape AI, make it work for us, 556 00:18:29,440 --> 00:18:30,480 create a better world. 557 00:18:30,480 --> 00:18:34,000 That's a great vision, something to strive for. 558 00:18:34,000 --> 00:18:35,000 But it won't be easy. 559 00:18:35,000 --> 00:18:37,000 There will be bumps along the way. 560 00:18:37,000 --> 00:18:38,720 Challenges, opportunities. 561 00:18:38,720 --> 00:18:41,600 We need to keep talking about AI, figuring it out together. 562 00:18:41,600 --> 00:18:44,600 I agree. The future of AI is not set in stone. 563 00:18:44,600 --> 00:18:47,000 We get to decide what it looks like through our choices, 564 00:18:47,000 --> 00:18:49,480 our actions. We need to make sure AI serves humanity, 565 00:18:49,480 --> 00:18:50,440 not the other way around. 566 00:18:50,440 --> 00:18:51,400 Well said. 567 00:18:51,400 --> 00:18:53,080 This has been a really interesting deep dive. 568 00:18:53,080 --> 00:18:56,800 We talked about everything, from robots escaping to lawsuits, 569 00:18:56,800 --> 00:19:00,520 the ethics of AI, even the meaning of life in an AI world. 570 00:19:00,520 --> 00:19:02,240 It's been a great discussion. 571 00:19:02,240 --> 00:19:05,160 And it shows why it's so important to stay informed, 572 00:19:05,160 --> 00:19:08,840 keep learning about AI, have these conversations 573 00:19:08,840 --> 00:19:11,040 so we can shape the future together. 574 00:19:11,040 --> 00:19:12,040 Thanks for joining us. 575 00:19:12,040 --> 00:19:14,680 Thanks for listening to the Daily AI News Podcast. 576 00:19:14,680 --> 00:19:43,640 And stay tuned for more.