Welcome to AI This Week, Gizmodo’s weekly deep dive on what’s been happening in artificial intelligence.
When I first saw a video of Dean Phillips, I thought it was a deepfake of Ron DeSantis. The guy’s head has roughly the same shape and hair color as the Florida governor’s, although his features are slightly less troll-like. Since then I’ve learned that no, actually, he’s a totally different guy. As a listless 2024 Democratic presidential candidate, Phillips still sorta resembles DeSantis, however.
To review, Phillips is currently running for President in the semi-fake Democratic primary against Joe Biden. Who is he? That’s a great question and it’s actually one that Phillips has been asked a lot lately. His X profile has a pinned tweet that reads: “If you’re asking “Dean Who?”, please allow me to introduce myself,” with a linked video that purports to answer the question. While there’s a lot about Phillips that is unknown, there are a few things we can say about him right off the top: He is currently serving as a U.S. Representative from Minnesota (and has been since 2018), he’s rich (in the 1990s, he inherited his stepfather’s liquor business and served as the CEO of the company from 2000-2012), and he’s a huge fan of AI. We’ll get to that last part in a minute.
As far as can be discerned, Phillips main argument for why he should be President is that he is not 80 years old. Indeed, he has been going around telling anybody who will listen that Joe Biden (who is 81) is too old to be President and that he’s going to lose the 2024 election to Donald Trump. By Phillips’ logic, he is not 81 (he’s 55), and, therefore, he’s qualified. However, if the candidate’s vitality is assured, what also seems clear is that he’s too smart to think that he’s actually going to be President during this election cycle. Newsweek notes that he “remains underwater by double-digit margins against Biden.” As such, all he’s really doing is making Joe Biden look bad.
Which all sorta begs the question: What is he really doing it for? The timeless answer (“to promote himself”) springs to mind, although there seems to be more to the story than that. At the same time that he’s running what feels like an obviously doomed campaign, Phillips has also been taking the opportunity to spread the gospel of Silicon Valley. Nice words about tech—most notably AI—have been a throughline of his campaign. Among other things, he has said that AI will be the “most transformational technology in human history” and has cautioned voters against being too quick to wish regulation upon the emergent industry. He recently went on former presidential hopeful Andrew Yang’s podcast and said nice words about AI before claiming that he was also interested in instituting universal basic income (UBI is the ploy tech bros have been trotting out as a way to justify taking away everybody’s jobs and giving them to robots). He has also said vaguely nice words about crypto, claiming at an event dubbed Stand with Crypto: “We should make sure we don’t stifle innovation, that we don’t stifle decentralisation when it is thoughtful and supportive of our national interest.” At a recent campaign event, he blurted out: “I will be the first AI president in American history!”
Why is Phillips pandering so hard to Silicon Valley? The mind boggles at the possibilities. It could be that he’s just trying to seem forward-thinking and modern next to his octogenarian opponent. It could be that he’s genuinely excited about AI and crypto and that he wants to share his earnest excitement with the world. Or, it could be that tech bros have been showering him with cash lately. Yeah, uh, that last one could potentially be it.
Indeed, a gaggle of tech denizens has been pouring cash into his campaign via a super-PAC called the We Deserve Better PAC. This PAC was actually formed by a former OpenAI employee, Matt Krisiloff, who has characterized his support for Phillips as an attempt to provide voters with an alternative to a candidate who is in his 80s and not exactly at the top of his game. In interviews with the press, Krisiloff has said that he believes there “should be open primaries” and “debates” and that people have noticed that Biden has been “slowing down very noticeably.” Phillips is also said to have had dinner with Krisiloff’s former boss, OpenAI CEO Sam Altman, towards the beginning of his campaign. He has since referred to Altman as “a source of great counsel, ideation, and perspective.”
Phillips also has ties (financially, at least) to billionaire Bill Ackman. Ackman, who made his fortune via a long career in financial investment and hedge fund management, is best known lately for his questionable interest in DEI initiatives at elite universities, plagiarism at Harvard, and Israel. But, like any savvy investor, Ackman also has a healthy interest in the monetary value of crypto and AI. Ackman previously lobbied for financial deregulation in New York to allow crypto to flourish there more readily; his hedge fund, Pershing Square Capital Management, has also invested heavily in Alphabet, Google’s parent company, apparently as a vote of confidence in the company’s deepening AI investments. Notably, Phillips’ We Deserve Better PAC has received as much as $1 million from Ackman, the Washington Post reports.
As a campaign gimmick, the We Deserve Better PAC recently launched a chatbot tailored to Phillips, dubbed the Dean.Bot. The bot was trained on audio and speeches made by the candidate and was designed to spit out answers when queried by interested voters. This probably scored some points with the tech-interested voters, although the bot ultimately had to be deactivated because OpenAI, whose algorithms were partially used in the chatbot, said the bot had violated its terms of service.
The support of the techie and billionaire class seems to have compelled Phillips to say some truly awful things, including that, should he win the presidency, his potential White House cabinet could include the likes of Ackman, Elon Musk, and techie angel investor Jason Calacanis. Holy shit, please don’t.
As previously stated, despite the ongoing infusions of cash into Phillips’s campaign he doesn’t have much of a shot at winning the presidency. Problematically, nobody knows who he is and, despite a better-than-expected showing in the New Hampshire primary, he is still woefully behind Biden in popular support. As one former Obama campaign manager, Jim Messina, recently said of wealthy donors’ financial support of Phillips: “It’s just an amazing way to piss away a million dollars.”
That said, the Phillips campaign does seem to have highlighted one particular political truth, which is that if money can’t buy a candidate a seat in the White House, what it can do is convert them into a living, breathing version of ChatGPT. Put more simply, if you feed them enough money, they’ll say pretty much whatever you want.
Question of the day: Can you tell the difference between an AI-generated face and a real human’s face?
The New York Times put out an interesting article this week that features a test to see whether readers can tell the difference between real human faces, and ones that have been generated using artificial intelligence. The Times story is based on recent research published by a host of researchers that argues we’ve arrived at the stage of “hyperrealism” wherein “AI-generated faces are now indistinguishable from human faces.” If my test results are any indication, those researchers appear to be right. I did not do very well and was only able to identify the correct answer about half of the time. If you want to test your bot-identifying skills, check out the Times story here.
More headlines this week
- The Bulletin of Concerned Scientists adds AI to our list of woes. For over 60 years, the Bulletin of Atomic Scientists has been counting down the days till we’re all toast. The group’s famous “Doomsday Clock” is meant as a symbolic barometer for how close we all are to being wiped out by existential threats—stuff like climate change, nuclear war, and now AI. This week the group conducted its annual ritual of resetting the Doomsday Clock, setting it at 90 seconds to midnight. According to the scientists, this is the “closest the Clock has ever been to midnight,” and, thus, the closest we—as a human species—have ever been to annihilation. Cool! Among the dangers that we all now face are: “Wars, Multi-Dimensional Nuclear Threats, Failures to Address the Climate Crisis, Bio-Threats, and Artificial Intelligence.” Someone, please let Sam Altman know ASAP.
- Mediocre software company crosses $3 trillion valuation, thanks to AI. This week, Microsoft’s stock price crossed the $3 trillion valuation mark, cementing its place as the second most valuable company in the world, next to Apple. The company, which is not generally known for being the best at anything except buying out competitors and ruthlessly cornering markets, seems to have gotten a pretty big boost from its corporate partnership with OpenAI and its aggressive expansion into artificial intelligence. That said, the high stock price doesn’t seem to have convinced the company’s C-suite not to fire a whole bunch of people. The company commenced with a massive round of layoffs this week, letting go as many as 1,900 people from its gaming division.
- Taylor Swift is reportedly considering legal action over AI-generated porn that features her. This week, some new AI porn made the rounds that happened to feature Taylor Swift. The porn, which featured Swift posing in suggestive and revealing ways, made its way onto X (formerly Twitter), where it picked up steam and was widely circulated. Now, Swift is apparently so pissed that she’s considering legal action.