Like the definition of an old man in a hurry, since his inauguration Donald Trump has been pumping out orders and vetoes like Lieutenant Kilgore summoning napalm strikes in Apocalypse Now. Such has been the shock-and-awe of his legislative blitz, some of Trump’s less newsworthy diktats have gone nearly unnoticed. But hidden away in one White House document, dated January 20 2025, and titled ‘INITIAL RESCISSIONS OF HARMFUL EXECUTIVE ORDERS AND ACTIONS’ is a policy switch that might save us all.
You have to scroll down the list of ‘harmful’ executive orders made by Biden – now being rolled back – to get to this jewel. But there it is – with one flourish of the Oval Office pen, Trump has wiped away Biden’s Executive Order 14110 of October 30, 2023 (Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence).
Why is this big? Surely an attempt to regulate this potent new technology and ensure its safety is a good thing? Certainly, if you read what Biden supposedly intended, it does look benign. It required developers of AI systems that ‘pose risks to US national security, the economy, public health or safety to share the results of safety tests with the US government, in line with the Defense Production Act’.
The order also instructed agencies to ‘set standards for [AI] testing and address related chemical, biological, radiological, nuclear, and cybersecurity risks’.
Which, again, sounds thoroughly desirable.
And yet many people in the AI industry (and in the Republican party) fought back bitterly against this government intrusion. Some saw it as an attack on free speech, a deliberate crippling of invention, and as an attempt to swaddle AI, in its cradle, within the dulling binds of wokeness. GOP Senator Ted Cruz said this in 2024:
‘The AI Executive Order… empowers federal bureaucrats to monitor the activities of all companies developing, or even considering developing, an AI model. Specifically, section 4 requires that wherever a company contemplates or develops a “dual-use foundation model”, it must submit to the government records of all AI training, development… describing how well the AI conforms with NIST’s woke AI ‘safety’ standards.’
Nor was this fear of a politically strangled AI some right-wing fever dream. Who can forget Google’s catastrophic Gemini AI (Google’s supposed competitor with ChatGPT), which was so marinated in wokeness, when asked to produce images of Nazis it made half of them black? Yes, after a backlash, Google retracted the model, but the concern lurked that Silicon Valley’s self-flagellating obsession with ‘correctness’ might chain this Promethean tech in the dungeons of do-goodery.
Now, the fear of an ideologically lamed AI – at the most crucial stage of AI development – has gone, at least in the USA. The EU is still keen on over-regulating AI – see its latest AI Act – but given that the EU’s AI industry consists of one man near Bordeaux trying to get his Amstrad to play boules, that is less concerning.
But there is an even more important side to this. Right now humanity – specifically in the USA and China – is locked in a race to achieve Artificial General Intelligence, then Artificial Superintelligence. Whoever gets to these first will possibly, probably, rule the world – or at least prevent others ruling. Even Biden officials admit the incredible salience of the moment.
A few days before Biden shuffled offstage, his security advisor, Jake Sullivan, called journalists to deliver a ‘chilling’ and ‘catastrophic’ warning for the incoming administration. Sullivan is quoted as saying, ‘Somehow, the government will have to join forces with [America’s AI] companies to nurture and protect America’s early AI edge, and shape the global rules for using potentially God-like powers.’
Before Trump dismissed Biden’s foolish Executive Order, there was a real risk that America’s tenuous but critical AI lead might be thrown away. There was talk of major AI companies shifting to the Middle East, where the Gulf States were offering limitless land and energy for those massive data centres. Now, that will not happen. Trump has just announced a $500 billion investment in AI over four years. In other words, this is a Manhattan Project for AI, a massive national undertaking to make sure the US gets there first.
And let’s hope the Americans succeed, and Trump saves the world. Because, whether you like or loathe Donald J Trump, the idea of Xi Jinping and China wielding ‘God-like’ hegemony over the rest of us is unthinkable.
Listen to the Americano podcast:
Comments