It’s somewhat amazing how the word “safety” was barely mentioned at this summit. It was more “How the world stopped worrying and learned to love AI”. Politicians have shifted from AI regulation to innovation — and the industry loves it.
12 February 2025 (Paris, France) – Having now retired from almost 25 years on the conference circuit – reporting to my readers from a most esoteric collection of events in art, cybersecurity, international affairs, media, politics, science and technology of all sorts – I now still attend a few, but more for fun and education.
One of my staffers attended Day 1 of the Paris conference, and we both did Day 2. She was more intent on covering themes. Me? I was more focused on minutiae. Although (damn it) I saw some “Big Picture” things, too, that I feel forced to share herein.
But as far as minutiae …
– “The Tale of the Missing Musk”. There were high hopes that Elon Musk would attend the AI summit in Paris, but late Monday evening, he RSVP’d “non”. It was a setback for French President Emmanuel Macron, hosting the summit, who is in regular touch with the tech mogul (3-4 calls a month, I am told). But they had a falling out – and its origins are a delicious dish. Musk was invited to and attended the historic ceremony reopening Notre Dame in December 2024, alongside world leaders, Musk insisted he would sit next to Trump, but Macron wanted Trump all to himself. Musk actually tried to sit next to Trump but was ultimately shunted back to the ninth row by ushers – quickly put in his place. So Musk was knocked back to sit behind French elected officials. Tant pis. Musk was very, very pissed.
– U.S. Vice President JD Vance, after making his keynote speech, left early – and so missed the speech of European Commission President Ursula von der Leyen, and even the closing speech by Macron. Sort of a “I’m really not interested in what you have to say” approach. And the evening before, at a dinner organized at the Elysée Palace with the business leaders present in Paris, he did not wait for dessert and slipped away, before the speech by Chinese Vice Premier Zhang Guoqing.
– However, earlier in the day Vance did meet with von der Leyen and High Representative Kaja Kallas (the EU’s version of a Foreign Secretary) on the margins of the summit. But the political feud between von der Leyen and Kallas was on full display. Kallas got airbrushed out. Von der Leyen mentioned her own name six times in her readout of the meeting — and didn’t note Kallas once. Von der Leyen also posted a photo of the meeting on X that showed no sign Kallas had been there, though she was sitting right next to her. She wants to show she is in charge of the U.S. relationship. In contrast, Kallas, to her credit, tagged the Commission chief in all of her social media posts, and included photos showing all of them meeting.
The focus of this event was clearly on Vance. He came, he spoke – and then he left, without listening to his counterparts’ speeches.
As I noted yesterday, he immediately asserted in his speech that “the United States of America is the leader in AI and our administration plans to keep it that way”. He then sounded the charge both against “censorship” by authoritarian regimes, targeting China without naming it, and against Europe’s desire to regulate these booming technologies. This was all in the name of defending “prosperous communities” and “free speech,” the unlimited freedom of expression so dear to the Trump administration.
And he multiplied his warnings. “Excessive regulation of the AI sector could kill a transformative industry,” he said, continuing that the Trump administration was also “troubled by reports that some foreign governments are considering tightening screws on US tech companies. America cannot and will not accept that”, tackling European regulations DSA (Digital Services Act), on social media, and GDPR (General Data Protection Regulation), on privacy. He finished with “We feel very strongly that AI must remain free from ideological bias and that American AI will not be co-opted into a tool for authoritarian censorship,” he said, referencing the moderation efforts required of American giants in Europe, but also in China.
The Trump administration’s crusade mirrors that of the American digital giants, who were very much in evidence in Paris, having for the most part pledged their allegiance to the billionaire president. “In this summit, the conversation has shifted from a discussion centered on the risks of AI, at the time of the London summit, to the idea that regulation must support innovation,” OpenAI’s global head of public affairs, Chris Lehane, said at a meeting on the sidelines of the summit, seeing an inflection in the massive announcements of investment in AI made by France and the EU. “European regulation, while well-intentioned, can actually be a risk to innovation,” said Lehane, deeming the European AI Act too “heavy-handed” compared to the Hiroshima process launched by the G7. The previous week, Meta had already announced that it would not sign the European code of good practice, judging it to be “impracticable.” OpenAI and its rival Anthropic also claim that “democracies” must win the AI race against “authoritarian regimes,” led by China.
The U.S. government and the U.S. tech oligarchs were all on the same page.
Faced with this American offensive, Macron, who had a long lunch with Vance on Tuesday, avoided any direct response. He deviated from his planned speech, and so he did not insist on European regulations, including the AI Act – which he once said was necessary to protect continental start-ups. He also tried to defend his idea of a “third way,” to “get out of the risk-opportunity dilemma,” while avoiding “regulation that could block innovation”. It did not fly very well amongst the crowd at the event.
And it has been clear that Europe wants to keep a low profile, with modest initiatives, and non-binding commitments.
When it was her turn to respond to Vance (who by this time was not in the room), von der Leyen tried to defend “Europe’s specific place for the global race in AI”. She said:
“Too often, I have heard that we should replicate what others are doing and run after their strengths. European-style AI should be based on collaboration between companies, sectors or countries, as well as on open AI”.
But on the subject of regulation, von der Leyen refrained from answering Vance’s question on “censorship.” And on the AI Act, she promised to simplify it by cutting “red tape.”
Faced with the U.S., the Europeans realized it is the only way to go. Petteri Orpo, the Finnish prime minister, who was also present in Paris, said in an interview with Le Monde:
“It’s a problem. As Europeans we need to find the right balance. JD Vance was very strong, but not overly aggressive. We are gong to need to somehow find the right balance between regulation and the dynamics of the AI sector, otherwise we risk losing out twice, because we won’t be able to convince our investors to stay on the continent and the Americans to come to us”.
But the big take-away for me was that safety fears, top of mind at all previous AI summits and talked about incessantly in the U.S., barely featured. The final summit declaration mentioned safety only three times. FYI: neither the United States or the U.K. signed the final declaration.
Instead, politicians touted massive investments and promised light regulation. As Vance himself said “I’m not here to talk about AI safety, I’m here to talk about AI opportunity”. Even von der Leyen seemed to join the chorus:
“This summit is focused on action, and that’s exactly what we need right now”.
But to be fair, this pivot from AI safety and governance to clinching deals and enabling AI companies was long in the making. France had purposely branded the event an “Action” summit, in stark contrast with the U.K.’s AI Safety Summit at Bletchley Park in November 2023, at which major tech companies committed to establishing safety frameworks. France reduced references to the Bletchley Summit to the minimum.
And look, timing also played a part, with a series of AI shocks in the weeks before the summit highlighting the global race between the U.S., China and the European Union:
• Trump’s $500 billion AI hardware plan
• Trumps cancellation of his predecessor’s AI safety rules, planting a flag for U.S. determination to win the AI race
• Markets erased billions of dollars from AI companies when the Chinese rival DeepSeek showed it could also develop AI models at a lower cost
Political realities in Brussels have also shifted. Von der Leyen is now focusing on unlocking growth for the region’s slowing economy. The deployment of AI is a top priority to achieve that goal. After the conference concluded, she posted on her Bluesy account “Europe is open for AI and for business, and still in the game!”
Vance and von der Leyen may actually be far more aligned on AI than we think. Where Vance called for rules that won’t strangle the burgeoning industry, von der Leyen promised the EU would cut red tape. She sent the EU’s tech sovereignty chief, Henna Virkkunen, around Paris, promising to make the EU’s regulatory framework more “innovation-friendly”. Virkkunen kept saying the EU was pledged to to simplify AI and tech rules as part of the EU’s efforts to simplify regulation. “We have too much bureaucracy and red tape”, she said.
A von der Leyen staffer said that does not mean the EU is abandoning its rules, with von der Leyen mentioning the new AI Act as a single set of safety rules for the bloc. But her announcement to free up tens of billions of euros for AI computing power sucked up most of the attention, and regulation was downplayed.
The politicians’ change of tone went down well with the AI industry, especially given the heavy regulatory scrutiny that some like chatbot pioneer OpenAI have faced in Europe. OpenAI executives highlighted the political pivot at a reception for reporters on the sidelines of the summit, even as they said safety issues still need to be addressed and that wider confidence in AI needs to grow. But the focus was “we must be willing to embrace this innovation because perhaps the biggest risk of all is actually missing out on the economic opportunities that come from this technology”. Not that they are just in it for the money, of course.
Obviously, it all negates the consensus from Bletchley. Digital rights groups saw the softer approach politicians were taking toward companies. The Brussels-based digital rights group EDRi which has been incredibly focused on AI “safety” and AI “ethics” slammed the EU’s decision to include the AI Act in a regulatory simplification push – something not envisioned until Trump rode into town. All they see now is deregulation, appeasement to the U.S. and its tech corporations, while ruining EU civil society – its few but hard-won human rights victories in the AI Act now dead.
Vance said plainly that the US sees AI as a race, that it plans to win that race, and that European-style regulation would represent an intolerable hindrance to innovation. He also warned that attempts to apply content moderation to AI systems represented authoritarian censorship. And several times he used the phrase “we believe that excessive regulation of the AI sector could kill a transformative industry”. He was laser focused on criticising the “massive regulations” created by the EU’s Digital Services Act, as well as Europe’s online privacy rules, known by the acronym GDPR, which he said meant endless legal compliance costs for smaller firms.
And perhaps hand-wringing about safety won’t “win the future,” if by “winning the future” you mean “building superintelligence before China does.” But Chinese dominance of AI is not the only or even the most serious risk of building machines that can out-think us — a fact that no amount of cowboy posturing can change.
But it does seems bizarre that at the highest levels of government, people believe that “we should just put super-powerful systems into the hands of as many people as possible and just see what happens”. Well, the tech oligarchs think it’s a good idea.
The brutal reality? Stuff I have written about before:
• Tech oligarchs were taking over the world
• They were overhyping LLMS as AGI, even as problems like hallucinations and boneheadedness still persist
• Governments are just too in bed with the companies, and will follow their lead
• We will wind up with little or no regulation around AI, despite massive risks around cybercrime, misinformation, bias, over-reliance (e.g., in military contexts) of inadequate systems, etc., etc., etc.
Vance came out swinging – exactly as the big companies might have hoped he might – that any regulation around AI was “excessive regulation” that would throttle innovation.
In reality, the phrase “excessive regulation” is sophistry. Of course in any domain there can be “excessive regulation”, by definition. What Vance doesn’t have is any evidence whatsoever that the US has excessive regulation around AI, Arguably, in fact, it has almost none at all. His warning about a bogeyman is a tip-off, however, for how all this is going to go. The new administration will do everything in its power to protect businesses, and nothing to protect individuals.
If AI has significant negative externalities upon the world, it’s that we, the citizens, are screwed.
And Vance took a veiled job at China, warning of authoritarian use of AI. He’s absolutely right, we should worry about authoritarian uses of AI.
But where exactly does he think that is going to lead – like, maybe, letting a private, AI-focused citizen virtually unlimited access to U.S. Treasury records, including some around personal finances? Also, having our own massive LLMs will do nothing to stop China from using theirs in authoritarian pursuits.
None of this brings me great joy. But this is where we are.