1 00:00:00,120 --> 00:00:00,880 Speaker 1: Mike Dubuski. 2 00:00:01,000 --> 00:00:06,400 Speaker 2: It is Tech Tuesday, ABC Technology Reporter and he's at 3 00:00:06,600 --> 00:00:11,319 Speaker 2: Michael Dobuski. That's d o b Uski. Mike, thank you 4 00:00:11,360 --> 00:00:13,760 Speaker 2: for joining us. Always appreciated. 5 00:00:14,440 --> 00:00:15,800 Speaker 3: Thanks for having me. Great to be here. 6 00:00:16,480 --> 00:00:22,919 Speaker 2: Yes, it is okay Trump. I love this EO on AI. 7 00:00:24,600 --> 00:00:27,080 Speaker 1: Okay, you want to explain that, because there's actually a 8 00:00:27,080 --> 00:00:29,240 Speaker 1: story there there is. 9 00:00:29,400 --> 00:00:32,720 Speaker 3: Yes, EO is Executive order and AI is of course 10 00:00:32,840 --> 00:00:35,360 Speaker 3: artificial intelligence, which we just can't seem to get away 11 00:00:35,360 --> 00:00:38,680 Speaker 3: from in the tech sector at this point. Two major 12 00:00:38,760 --> 00:00:41,400 Speaker 3: executive orders kind of in the mix here in the 13 00:00:41,440 --> 00:00:45,600 Speaker 3: Trump administration this week. Yesterday, on Monday, the Trump administration 14 00:00:45,800 --> 00:00:49,559 Speaker 3: signed what they call Project Genesis, which is basically an 15 00:00:49,560 --> 00:00:53,680 Speaker 3: executive order designed to spur the use of generative artificial 16 00:00:53,720 --> 00:00:57,040 Speaker 3: intelligence in scientific research. In fact, it directs the Department 17 00:00:57,080 --> 00:01:00,360 Speaker 3: of Energy to actually build its own AI plant form 18 00:01:00,800 --> 00:01:04,880 Speaker 3: fed by data collected by the federal government that academics 19 00:01:04,920 --> 00:01:07,760 Speaker 3: can use in their research. So it's another tool in 20 00:01:07,800 --> 00:01:10,800 Speaker 3: their toolbox. That was signed by the President on Monday, 21 00:01:10,840 --> 00:01:13,640 Speaker 3: But the bigger executive order could very well come later 22 00:01:13,680 --> 00:01:15,680 Speaker 3: this week, and this has been generating a lot of 23 00:01:15,720 --> 00:01:20,160 Speaker 3: conversation in Washington to The Trump administration is reportedly considering 24 00:01:20,480 --> 00:01:24,160 Speaker 3: this EO that would put a halt to AI legislation 25 00:01:24,720 --> 00:01:28,119 Speaker 3: at the state level. The Trump administration the goal here, says, 26 00:01:28,160 --> 00:01:30,320 Speaker 3: The goal here is to make it easier for AI 27 00:01:30,400 --> 00:01:34,000 Speaker 3: companies to develop this technology a mid what they see 28 00:01:34,000 --> 00:01:37,560 Speaker 3: as a global race for AI supremacy. They say that 29 00:01:37,720 --> 00:01:40,880 Speaker 3: clearing the field of a patchwork of AI legislation at 30 00:01:40,880 --> 00:01:44,600 Speaker 3: the state level and in favor of sort of forthcoming 31 00:01:44,680 --> 00:01:47,319 Speaker 3: AI legislation at the federal level will make it a 32 00:01:47,319 --> 00:01:50,040 Speaker 3: lot easier for companies like open ai and Google and 33 00:01:50,080 --> 00:01:53,480 Speaker 3: Anthropic to navigate as they build large language models to 34 00:01:53,520 --> 00:01:55,800 Speaker 3: be competitive with specifically China. 35 00:01:56,360 --> 00:01:57,920 Speaker 1: Okay, so let me let me ask this. 36 00:01:58,000 --> 00:02:02,800 Speaker 2: Okay, we know that the denial of money to universities 37 00:02:02,800 --> 00:02:06,320 Speaker 2: for medical research, et cetera, that is a political move. 38 00:02:06,400 --> 00:02:09,040 Speaker 2: I mean, no one disagrees, whether you're one side or 39 00:02:09,080 --> 00:02:12,320 Speaker 2: the other, No one disagrees that that the basis of 40 00:02:12,560 --> 00:02:18,520 Speaker 2: that policy is political. Moving over to this executive order, 41 00:02:19,280 --> 00:02:21,720 Speaker 2: can this is this political? 42 00:02:21,760 --> 00:02:23,519 Speaker 1: Does it have any political overtones? 43 00:02:25,120 --> 00:02:26,680 Speaker 3: Which one are we talking about? The one that's already 44 00:02:26,680 --> 00:02:28,919 Speaker 3: been signed or the one that has not been signed yet? 45 00:02:29,120 --> 00:02:33,360 Speaker 3: Both both well, I mean, I think, you know, anything 46 00:02:33,360 --> 00:02:37,520 Speaker 3: that comes from really any presidential administration is by nature political. 47 00:02:38,000 --> 00:02:40,840 Speaker 3: The one to sort of spur scientific research, you know, 48 00:02:40,919 --> 00:02:43,880 Speaker 3: seems pretty standard. It doesn't seem to be generating a 49 00:02:43,919 --> 00:02:46,919 Speaker 3: ton of controversy out there, but certainly the one that 50 00:02:47,400 --> 00:02:51,000 Speaker 3: you know is forthcoming that reportedly, you know, the Trump 51 00:02:51,000 --> 00:02:55,760 Speaker 3: administration is working on to curtail state level AI legislation 52 00:02:56,040 --> 00:02:59,560 Speaker 3: is really generating some some heated political back and forth, 53 00:02:59,639 --> 00:03:02,560 Speaker 3: and interestingly enough, it comes from both sides of the aisle. 54 00:03:02,560 --> 00:03:06,040 Speaker 3: It's kind of the Trump administration against state governments. For example, 55 00:03:06,040 --> 00:03:08,959 Speaker 3: we've seen Florida Governor Ron DeSantis, a Republican, of course, 56 00:03:09,040 --> 00:03:13,080 Speaker 3: ran against President Trump in the Republican primary last cycle. 57 00:03:13,400 --> 00:03:16,720 Speaker 3: He called this move federal government overreach that would stop 58 00:03:16,760 --> 00:03:20,760 Speaker 3: state's ability to protect against what he calls predatory applications, 59 00:03:20,760 --> 00:03:23,840 Speaker 3: the target children, and of course, Democratic Senator Ed Markey 60 00:03:23,919 --> 00:03:26,800 Speaker 3: says this move would amount to President Trump siding with 61 00:03:26,919 --> 00:03:30,160 Speaker 3: his billionaire buddies in Silicon Valley. There have also been 62 00:03:30,200 --> 00:03:33,360 Speaker 3: several high profile letters to Congress aimed at raising the 63 00:03:33,400 --> 00:03:37,280 Speaker 3: alarm about this forthcoming executive order. The Leadership Conference Center 64 00:03:37,320 --> 00:03:40,880 Speaker 3: for Civil Rights and Technology says this gives corporations a 65 00:03:40,960 --> 00:03:44,480 Speaker 3: free pass rather than protecting the people it's meant to serve. 66 00:03:44,760 --> 00:03:47,120 Speaker 3: And of course there is some reporting out there, namely 67 00:03:47,120 --> 00:03:49,920 Speaker 3: from Reuters, which indicates that this might actually be working 68 00:03:50,040 --> 00:03:52,760 Speaker 3: on the Trump administration then reporting that they might be 69 00:03:52,800 --> 00:03:55,680 Speaker 3: adjusting their timeline. We originally thought that this order was 70 00:03:55,920 --> 00:03:59,160 Speaker 3: going to be signed this week. That might be sliding somewhat. 71 00:03:59,200 --> 00:04:02,120 Speaker 3: And of course, I forget earlier this year during the 72 00:04:02,160 --> 00:04:04,760 Speaker 3: passage of the defensele to some of the risks associated 73 00:04:04,800 --> 00:04:07,000 Speaker 3: with AI. It's going to be really interesting to see 74 00:04:07,000 --> 00:04:09,600 Speaker 3: how this plays out because the Trump administration needs to 75 00:04:09,600 --> 00:04:12,640 Speaker 3: balance both the sort of national and global race for 76 00:04:12,800 --> 00:04:15,560 Speaker 3: AI with a lot of the satan concerns inherent to 77 00:04:15,640 --> 00:04:16,360 Speaker 3: this technology. 78 00:04:16,640 --> 00:04:20,119 Speaker 2: Mike, we're talking about what the Trump administration. What Donald 79 00:04:20,120 --> 00:04:24,280 Speaker 2: Trump did was sign on to well, he sign an 80 00:04:24,279 --> 00:04:29,000 Speaker 2: executive order on taking AI and wants to take it federally. 81 00:04:29,400 --> 00:04:30,400 Speaker 1: In terms of the laws. 82 00:04:31,160 --> 00:04:35,599 Speaker 2: Now, having been in reproductive law, my entire legal career 83 00:04:35,720 --> 00:04:41,480 Speaker 2: was third party reproductive law. The technology was always miles 84 00:04:41,560 --> 00:04:44,200 Speaker 2: ahead of the law. The law never keeps up with 85 00:04:44,240 --> 00:04:50,320 Speaker 2: the technology. How aggressive are the states and the federal government? 86 00:04:50,400 --> 00:04:52,320 Speaker 1: And going after controlling. 87 00:04:51,839 --> 00:04:56,080 Speaker 3: This again, it really does seem to be a state 88 00:04:56,320 --> 00:05:00,120 Speaker 3: effort at this point. There is no federal regulation governing 89 00:05:00,240 --> 00:05:03,400 Speaker 3: the use or deployment of AI. That is really what 90 00:05:03,560 --> 00:05:07,320 Speaker 3: spurred a lot of these states, among them California, Colorado, Utah, 91 00:05:07,520 --> 00:05:12,480 Speaker 3: Texas to pass legislation at their level. So, yeah, you're 92 00:05:12,480 --> 00:05:15,400 Speaker 3: absolutely right to say that the pace of technological development 93 00:05:15,440 --> 00:05:19,960 Speaker 3: does outpace the level at which we can pass you know, 94 00:05:20,040 --> 00:05:22,719 Speaker 3: substantive laws. But it does seem like we're at a 95 00:05:22,760 --> 00:05:26,560 Speaker 3: stage now where there's this real consideration among lawmakers as 96 00:05:26,600 --> 00:05:29,920 Speaker 3: to who's going to go about this right if not Congress. Well, 97 00:05:30,520 --> 00:05:33,800 Speaker 3: certain states want some protections in place, and just to 98 00:05:33,800 --> 00:05:36,479 Speaker 3: give you a little overview of what those laws look like, 99 00:05:36,839 --> 00:05:40,360 Speaker 3: you know, in California, you guys have passed laws AI 100 00:05:40,480 --> 00:05:44,479 Speaker 3: legislation aimed at limiting the collection of personal data from 101 00:05:44,520 --> 00:05:47,800 Speaker 3: AI companies, forcing AI companies to be more transparent with 102 00:05:47,880 --> 00:05:50,239 Speaker 3: the data they do collect. Of course, that's very relevant 103 00:05:50,279 --> 00:05:53,640 Speaker 3: given how many tech companies are based in California, but 104 00:05:53,640 --> 00:05:56,560 Speaker 3: we've also seen that, as we said, in Colorado, Utah, Texas, 105 00:05:56,600 --> 00:06:00,000 Speaker 3: other states beyond that have passed regulations around using AI 106 00:06:00,200 --> 00:06:03,919 Speaker 3: to generate deep fakes of political figures and using AI 107 00:06:04,040 --> 00:06:07,479 Speaker 3: to create non consensual pornography, so it really does cover 108 00:06:07,520 --> 00:06:10,640 Speaker 3: a broad remit. But if this executive order does go 109 00:06:10,800 --> 00:06:13,880 Speaker 3: into effect, this would be basically clean the slate, wipe 110 00:06:13,920 --> 00:06:17,600 Speaker 3: the table in uh, you know, anticipation of some future 111 00:06:17,640 --> 00:06:18,920 Speaker 3: federal regulation to come. 112 00:06:19,400 --> 00:06:19,560 Speaker 1: Right. 113 00:06:19,600 --> 00:06:23,000 Speaker 2: And one last question about that, and that is is 114 00:06:23,040 --> 00:06:27,320 Speaker 2: this front front burner stuff for the states that they're 115 00:06:27,400 --> 00:06:29,479 Speaker 2: viewing this as absolutely critical and we have to do 116 00:06:29,520 --> 00:06:32,599 Speaker 2: it now, I would say increasingly. 117 00:06:32,680 --> 00:06:35,599 Speaker 3: So, yeah, absolutely, given you know how many stories have 118 00:06:35,640 --> 00:06:38,279 Speaker 3: been emerging about you know, AI going wrong for people. 119 00:06:38,560 --> 00:06:41,160 Speaker 3: You know this that's an unfortunate story, but you know 120 00:06:41,200 --> 00:06:44,120 Speaker 3: many people have you know, become obsessed with many of 121 00:06:44,120 --> 00:06:46,960 Speaker 3: these chatbots. They are in some small way designed to 122 00:06:47,440 --> 00:06:50,120 Speaker 3: you know, uh, support your thinking and to keep you 123 00:06:50,240 --> 00:06:53,400 Speaker 3: using them. That is true of not to this AI platforms, 124 00:06:53,440 --> 00:06:56,760 Speaker 3: many tech platforms as well. People have even you know, 125 00:06:56,800 --> 00:06:59,719 Speaker 3: cut off communication with their families, gone missing, and some 126 00:07:00,080 --> 00:07:02,039 Speaker 3: dream case has actually taken their own lives and the 127 00:07:02,080 --> 00:07:04,480 Speaker 3: lives of others. And that really does get the attention 128 00:07:04,520 --> 00:07:06,719 Speaker 3: of lawmakers, you know, who say, hey, we got to 129 00:07:06,760 --> 00:07:09,640 Speaker 3: do something. You know, the federal government's not going to act, 130 00:07:09,680 --> 00:07:11,320 Speaker 3: we at state level will all. 131 00:07:11,280 --> 00:07:17,080 Speaker 2: Right, switching gears story about the genesis of an X account. 132 00:07:17,120 --> 00:07:18,840 Speaker 1: Where's your ex account coming from? 133 00:07:19,240 --> 00:07:22,200 Speaker 2: This broke and this actually turns out to be a 134 00:07:22,240 --> 00:07:24,440 Speaker 2: huge political issue on top of everything else. 135 00:07:25,840 --> 00:07:28,400 Speaker 3: Yeah, it does. And this new feature that X rolled 136 00:07:28,400 --> 00:07:31,120 Speaker 3: out over the weekend, the company says, is designed to 137 00:07:31,240 --> 00:07:34,960 Speaker 3: provide more transparency into people's accounts, basically, so you know 138 00:07:35,040 --> 00:07:38,480 Speaker 3: who you're interacting with online. They've gone about this in 139 00:07:38,480 --> 00:07:40,520 Speaker 3: a couple different ways. You've already been able to see 140 00:07:40,560 --> 00:07:43,200 Speaker 3: when someone signed up for the platform. It tells you 141 00:07:43,280 --> 00:07:46,240 Speaker 3: kind of what they joined or the general time period 142 00:07:46,240 --> 00:07:48,440 Speaker 3: in which they joined. You can now see how many 143 00:07:48,480 --> 00:07:51,720 Speaker 3: times a given account has changed their username, which could 144 00:07:51,720 --> 00:07:54,080 Speaker 3: be interesting. But the big example is that it will 145 00:07:54,120 --> 00:07:57,560 Speaker 3: now show where that account is based. And the really 146 00:07:57,560 --> 00:07:59,280 Speaker 3: big thing that's been happening as a result of this 147 00:07:59,360 --> 00:08:01,200 Speaker 3: feature is that's some people have been discovering that the 148 00:08:01,240 --> 00:08:04,480 Speaker 3: accounts they've been following, specifically the ones that cover politics, 149 00:08:04,720 --> 00:08:07,960 Speaker 3: actually originate from outside of the United States. Just some 150 00:08:08,080 --> 00:08:10,920 Speaker 3: examples here. There's one account with over two hundred thousand 151 00:08:10,920 --> 00:08:15,000 Speaker 3: followers last week called American Voice. It posted you know, 152 00:08:15,080 --> 00:08:18,480 Speaker 3: updates about the news and you know, various sort of 153 00:08:18,600 --> 00:08:20,400 Speaker 3: you know, tweets and posts about what was going on 154 00:08:20,400 --> 00:08:23,280 Speaker 3: in American politics. After this update rolled out, followers found 155 00:08:23,280 --> 00:08:25,560 Speaker 3: it was being run from somewhere in South Asia, and 156 00:08:25,600 --> 00:08:28,280 Speaker 3: now that account has been deleted. There was another account 157 00:08:28,320 --> 00:08:31,200 Speaker 3: that posted about President Trump's daughter, just news items about 158 00:08:31,200 --> 00:08:34,040 Speaker 3: Avanka Trump that according to this feature, that account is 159 00:08:34,080 --> 00:08:38,480 Speaker 3: actually based in Nigeria, And an account called Abujama Gaza that, 160 00:08:38,600 --> 00:08:42,120 Speaker 3: according to this account's profile, is a Gaza based journalist 161 00:08:42,200 --> 00:08:45,199 Speaker 3: covering the war. According to this feature, it's actually based 162 00:08:45,200 --> 00:08:48,199 Speaker 3: in Poland. So it really does underscore how the information 163 00:08:48,800 --> 00:08:51,920 Speaker 3: environment on access maybe a little bit more complex than 164 00:08:51,920 --> 00:08:52,800 Speaker 3: we initially thought. 165 00:08:53,080 --> 00:08:56,079 Speaker 2: Okay, so one quick question before I bail. In terms 166 00:08:56,120 --> 00:09:00,600 Speaker 2: of the political side of this, are these daw x 167 00:09:00,640 --> 00:09:06,640 Speaker 2: accounts more conservative, more liberal, more conspiracy theorists, or is 168 00:09:06,679 --> 00:09:07,480 Speaker 2: across the board. 169 00:09:08,320 --> 00:09:11,200 Speaker 3: So we don't have a ton of hard numbers on this, 170 00:09:11,400 --> 00:09:13,880 Speaker 3: but it is worth saying that these accounts do come 171 00:09:13,880 --> 00:09:18,359 Speaker 3: from both sides of the aisle. They are not exclusively 172 00:09:18,400 --> 00:09:21,000 Speaker 3: political accounts. Sometimes they're just accounts that post sort of 173 00:09:21,080 --> 00:09:24,760 Speaker 3: viral content or content that goes wide on that platform. 174 00:09:24,840 --> 00:09:28,120 Speaker 3: The bigger picture here really is that since Elon Musk 175 00:09:28,160 --> 00:09:30,520 Speaker 3: took over the platform, there's been really an intense effort 176 00:09:30,600 --> 00:09:33,480 Speaker 3: to monetize viral posts. So if you post something that 177 00:09:33,520 --> 00:09:35,520 Speaker 3: a lot of people like or reshare, you can make 178 00:09:35,520 --> 00:09:37,160 Speaker 3: a couple bucks off that, And to you or me, 179 00:09:37,360 --> 00:09:40,000 Speaker 3: that's just a couple bucks, it's not really that much. 180 00:09:40,080 --> 00:09:43,480 Speaker 3: But in developing countries specifically, that can actually amount to 181 00:09:43,559 --> 00:09:46,640 Speaker 3: quite a lot of money. So there's this incentive structure 182 00:09:46,720 --> 00:09:49,760 Speaker 3: at X to create viral content. I think the reason 183 00:09:49,800 --> 00:09:51,720 Speaker 3: that we're talking about in the context of politics is 184 00:09:51,720 --> 00:09:54,439 Speaker 3: that when you post something about politics on social media, 185 00:09:54,600 --> 00:09:57,160 Speaker 3: generally speaking, that does get a lot of engagement. Not 186 00:09:57,200 --> 00:10:01,000 Speaker 3: always get good engagement, but any engagement does equal dollars, 187 00:10:01,120 --> 00:10:03,240 Speaker 3: and I think that's exactly why you're seeing this happen 188 00:10:03,320 --> 00:10:03,680 Speaker 3: right now. 189 00:10:04,160 --> 00:10:05,680 Speaker 1: Is Musk making money on X? 190 00:10:05,720 --> 00:10:09,120 Speaker 2: I know he overpaid for it ridiculously, but has it 191 00:10:09,160 --> 00:10:12,120 Speaker 2: turned around and monetize enough to where Lisa is not 192 00:10:12,280 --> 00:10:13,199 Speaker 2: burning through money? 193 00:10:14,400 --> 00:10:16,360 Speaker 3: It's an interesting question. Again, we don't have a ton 194 00:10:16,360 --> 00:10:18,600 Speaker 3: of insight into his personal finances when it comes to 195 00:10:18,640 --> 00:10:21,840 Speaker 3: this company, although you're correct to say that he did 196 00:10:21,880 --> 00:10:25,280 Speaker 3: overpay by quite a hefty margin, the tens of billions. 197 00:10:25,320 --> 00:10:27,960 Speaker 3: When it comes to purchasing this platform back in twenty 198 00:10:28,000 --> 00:10:31,600 Speaker 3: twenty three. But another sort of interesting data point that 199 00:10:31,600 --> 00:10:33,720 Speaker 3: people might not be paying attention to is that despite 200 00:10:33,760 --> 00:10:35,600 Speaker 3: the fact that we talked about how the changes that 201 00:10:35,640 --> 00:10:39,440 Speaker 3: Elon Musk made to X drove this exodus from the platform, 202 00:10:39,440 --> 00:10:42,240 Speaker 3: a lot of people stopping using that platform because they 203 00:10:42,240 --> 00:10:44,280 Speaker 3: didn't agree with the policies of Elon Musk, a lot 204 00:10:44,320 --> 00:10:47,400 Speaker 3: of advertisers leaving a lot of those advertisers have actually 205 00:10:47,400 --> 00:10:49,960 Speaker 3: come back in the years since, and a recent Pew 206 00:10:50,080 --> 00:10:52,840 Speaker 3: survey found that more than one in five Americans still 207 00:10:52,960 --> 00:10:56,000 Speaker 3: use X formerly Twitter. While the competitors that we talked 208 00:10:56,000 --> 00:10:58,960 Speaker 3: about a lot, things like Threads and Blue Sky, they 209 00:10:59,000 --> 00:11:03,640 Speaker 3: remain comparabletively bit players in this space. So despite all 210 00:11:03,679 --> 00:11:07,160 Speaker 3: the talk about it, they do remain a pretty hasty, 211 00:11:07,360 --> 00:11:09,960 Speaker 3: you know, force in the world of social media. 212 00:11:10,080 --> 00:11:13,120 Speaker 2: Mike, thank you, we'll talk again. Always, great stuff. I 213 00:11:13,200 --> 00:11:16,559 Speaker 2: have a good day, you too, take care, Happy Thanksgiving. 214 00:11:17,160 --> 00:11:21,000 Speaker 2: That's what you're supposed to say, isn't it. That's what 215 00:11:21,080 --> 00:11:21,560 Speaker 2: they tell me.