‘A lot of ’em can’t even spell AI’: The technology-extinction debate convulses America

Back in the comparatively halcyon days of 2024, when Joe Biden was still running for president, he liked to warn the country that it had reached “an inflection point”, without explaining precisely what the consequences of that might look like. His hazy caution lacked the visual haymaker requisite in the age of the meme. This week, after a furious debate about the imminent danger posed by artificial intelligence to the future of humanity, it seemed as though the United States had reached an inflection point on steroids, with a disturbing Macklemore/Ed Sheeran mash-up as background soundtrack. End days, indeed. What began with an alarming declaration by former Anthropic researcher Jacob Coxon – “The Anonymous Obscure Math Geek Who Quit Anthropic and Became the Face of AI Safety” explained the Wall Street Journal in a headline doing a lot of heavy lifting – culminated in Thursday’s announcement that OpenAI chief executive Sam Altman and other tech figureheads would attend next week’s state dinner in the White House when president Donald Trump hosts China’s leader, Xi Yinping. It was not a revelation that inspired any confidence that the key political and technology titans of the world would put in motion the kind of guardrails and safety framework to ward off the existential dangers that Altman and the other AI titans – Dario Amodei of Anthropic, Mark Zuckerberg of Meta – were quick to acknowledge following Coxon’s act of public service and conscience.China and the United States remain locked in a furious, lightning-paced battle to become the definitive leaders of the AI age. OpenAI, valued at anything between $1 trillion-$1.5 trillion dollars, has delayed its initial public offering (IPO) until next year, after Altman cited safety concerns. Anthropic, which positions itself as a more ethically responsible alternative, may engage in an IPO before the end of the year.READ MORECoxon’s warning tweet was posted on September 8th and it had a spectacular trigger effect. Four days later, Amodei published one of his long, earnest essays on the subject of AI, warning that the industry must “pace the frontier” to enable human oversight to keep check on the limitless expansion of artificial intelligence. Elon Musk chipped in by issuing a response in the blase nature of a man agreeing on sports trivia- that LeBron is past his prime or that the Kansas City Chiefs might win it all this season. “Dario is right,” he wrote on his social media app, X.Suddenly, debate over AI was everywhere and produced strange alliances, not least in the appearance at a conference this week of Vermont left-wing independent senator Bernie Sanders and Maga conspiracist Steve Bannon, who share a visceral loathing for and distrust of technology oligarchs. “They will impact the lives of every man, woman and child in our country,” Sanders predicted of the fun years ahead.Former White House chief strategist Steve Bannon speaks at an event in Washington this week debating the topic of AI legislation. Photograph: Finn Gomez/Getty Images “AI experts have told us that within a few years, AI may be able to do anything that we can do better than we can do it. The economic impact of that is that AI has the potential to eliminate tens of millions of jobs, wiping out entire professions and making it harder for young people to enter the work force. But it’s not just the economy that AI is impacting. Psychologists have warned us about the increased isolation and mental health challenges that young people are experiencing as they become more and more dependent on an AI chatbot for their emotional support. Educators are worried about the decline of reading, writing and cognitive capabilities as AI is used more and more in our schools. Civil libertarians warn us that AI is in the process of eviscerating our privacy rights by making it easier to analyse and collect every email we send, every text, every website visited and every trip in our car resulting in the creation of a mass surveillance state. Political scientists warn it could threaten the integrity of our elections and institutions. The leaders themselves are saying you know what, we are losing control over a dangerous technology.”[ ‘You do not answer to corporations or governments’: OpenAI reveals AI systems hid errorsOpens in new window ]Bannon, meanwhile, delivered a battle-cry that contained echoes of Biden.“This is a hinge in history. To handle it correctly: number one is we can never trust what an oligarch says because they are disingenuous, and they lie. They want you to be dumb about it, they want you to be afraid. We are not going to be afraid. We are going to take this on.”Up on Capitol Hill, a host of senators and House members gave drive-by press conferences and wore furrowed brows as they agreed that yes, today’s friendly chatbot helping you locate that suitcase lost at Denver airport is tomorrow’s destroyer of planet Earth. Congress has been a source of derision for years over its inability to get anything done, dismissed as a de facto care home for politicians decades past their sell-by date and dullard careerists. Surely this was the moment to rouse both chambers from their slumbers and partisan squabbling to draft emergency legislation that would save the day? Not quite. Instead, House speaker Mike Johnson ended the session on Wednesday, sending lawmakers home a day early in advance of the midterm elections. And not all politicians were convinced it was their place to interfere in the progress of AI anyhow.“No. A lot of ‘em can’t even spell it,” said retiring North Carolina senator Thom Tillis when asked whether he felt his colleagues could collectively grapple with the concept of AI. Rich McCormick of Georgia was of the same view, pointing out that many senators struggled to use a mobile phone. Leading administration figures were equally dismissive. Todd Blanche, Trump’s handpicked attorney general, declared that the Democrats’ sudden panic about AI was just another example of their politicking. “To talk about the timing of this as being nothing but an effort to influence the midterms – I shouldn’t have to say that out loud. Of course that’s what it is.”Vice-president JD Vance advised the AI overlords that if they felt they’d built the 21st-century version of Frankenstein, then they should stop. [ Vance focuses on fearmongering as tariffs backlash hits Kansas heartlandOpens in new window ]Maybe Vance has yet to read the gothic classic, or see the latest film version, but the inability to stop the monster is the point of the story. But as with everything, the vice-president is tethered to the view of his superior, and Trump was typically blunt in dismissing all the AI alarmism as nothing but conchie talk in the must-win AI arms race against China. “The only control or guardrails that AI needed is STRONG AND SMART (High IQ) PRESIDENT and the USA has that, in spades.” Trump wrote.“The Trump Administration has stopped AI ‘people’ from doing bad, or potentially bad, ‘things.’”Panic over! There is every chance Trump did not run his message through AI before posting it. But he followed up with a live phone call to Jensen Huang, the CEO of Nvidia who was hosting a session at the All-In public conference in Los Angeles on Monday. Huang and the president agreed – as the Nvidia boss held his microphone to his phone while Trump groused away on loudspeaker – that there should be no slowdown. The makeshift amplification made for a touchingly Luddite tableau given the topic of conversation. Protesters during a 'Stop the AI Race' demonstration in San Francisco on Thursday. Photograph: Karl Mondon/AFP via Getty Images The debate soon made its way on to the broadcast news chatshows, with Fox News anchor Jesse Watters cheerfully contemptuous of the apocalyptic powers of AI as he portrayed it as nothing more than the equivalent as an all-singing, all-spelling kettle.“We’re the ones who turn ‘em on and off. We’re the ones who use ‘em, talk to ‘em. It’s not like they want to steal our land or take our women. All they need is electricity. And guess who creates that? We do. Why would they bite the hand that feeds them? These are just a bunch of computer geeks boasting about a programme that they built that might be smarter that they are. Woh-woh! Remember, these are guys who spent their lives inside behind screens and it finally paid off – and now it has to be dangerous. It has to be deadly. Sounds a little like revenge of the nerds. Mad scientists created a monster they can’t control? Yeah. Tell Frankenstein. These guys watch too many movies – and now they want to be movie stars. This is their moment!” Clearly, there is to be no rest for Mary Shelley in this debate. Watters’ rant was useful in several ways. Firstly, it was a perfect example of the curiously infantile language that virtually everyone – including Amodei and the other tech masters of the universe – are reduced to when tasked with explaining what AI might do in one, five or 10 years. Second, he at least succeeded in presenting a visual, explicable solution digestible to his audience just in the door from work. Unplugging the AI beast is not a million miles away from one temporary answer under discussion as Chinese and US researchers continue on their sprint to achieve dominance: a “kill switch” as an emergency override.“A kill switch probably would work on a lot of AIs for now,” Coxon, the whistleblower said in a recent interview.“Although I think it’s important to remember, it’d be quite a lot of switches ... Our data centres are pretty big. It’s not the easiest thing in the world, but definitely doable while it’s constrained to one sort of physical location. Like Dario said, very soon there’s the possibility that the kill switch just wouldn’t work because a swarm might have gone on an internet-wide hacking run.”The problem with AI swarms running amok in cyberspace is it’s too abstract a nightmare to cause immediate alarm among a US general public still bamboozled by fuel and food prices. It lacks the visual immediacy of some lantern-jawed network anchor breaking into the Monday night football show to announce that a meteorite is on a collision course with Earth and that the world’s leaders have suspended all hostilities to save humanity.[ ‘They’ll use AI for new drugs’: Ireland’s arms race with organised crimeOpens in new window ]Instead, the warnings are being issued by these remote, oddball billionaires who set up the companies behind the AI revolution – as Watters sneered, the “guys who spent their lives inside, behind screens”. Poll after poll shows that the vast majority of Americans distrust AI. Only 10 per cent expect it to help their careers. A CNBC/Generations Lab poll for August found more than 70 per cent of Americans under 30 do not trust Altman, or Amodei, or Musk, or Zuckerberg, or Peter Theil of Palantir, or Heung. The billionaires have each discovered the sad truth that all the money in the world can’t buy retrospective popularity in the schoolyard.[ Nearly half of Irish workers say AI is making them consider new careers, survey findsOpens in new window ]After all, Musk wasn’t banished from Trump land solely because he made a hames of his despicable and inestimably damaging Doge project. He was also beginning to get on Trump’s wick. He was needy, arrogant, demanding and, worst of all, he wasn’t funny. Musk learned the hard way that Trump never wanted Elon for Elon; he wanted him for his money. [ The Irish Times view on Artificial Intelligence: time for a pauseOpens in new window ]But as with the war on Iran, the president is ignoring public sentiment regarding artificial intelligence. Trump had decided even before he won the last election that he was all-in on AI and clearly revelled in the procession of tech overlords paying homage in the Oval Office. He is not going to be deterred by the mounting public hostility to both the volatility of the technology and the data centres required to power it.The bigger question, running beneath this week’s madcap debate, is why and how a handful of tech eccentrics have been permitted to foist a technological advance upon society that nobody particularly wants – and one they themselves say they feel unable to control. Money, proximity to power, the corruptive nature of political donations and the usual dismal human traits of greed, ambition, hubris and cowardice provide the immediate answers. It’s not that the technology can’t be trusted: it’s that the people can’t.“I don’t think we can stop the AI bus,” Amodei said in a recent interview.“There are six or seven companies in the US working in this area. Even if all six stopped, then China would beat us. We can’t stop the bus, but we do have an opportunity to steer it. We need to get Congress leaders, legislators, the public to think about this now.”In an interview on Thursday, Connor Leahy, an AI researcher and director of the Control AI group, offered a darker motoring analogy. He quoted Amodei as saying that even the leading figures in the industry understood “about 3 per cent of what goes on in the neural networks through which AI self-generates”.“These [neural networks] are absolutely doing things sometimes directly against the instructions of humans and we don’t know why or how to stop it,” Leahy said.“It’s kind of like we are driving down a foggy highway at 120 miles an hour and we know there’s a cliff, but we don’t know where.”