On Wednesday in London, the tenth leg of his tour on five continents, Sam Altman hardened the tone by warning that OpenAI could "stop operating" in the European Union if the future European regulation imposed too many constraints.

"We will try (to comply) but there are technical limits to what is possible," he told the British press, explaining that he had "a lot" of criticism of the EU's future AI Act. This triggered the anger of European Commissioner Thierry Breton on Thursday.

"There is no point in trying to blackmail by claiming that by developing a clear framework, Europe is delaying the deployment of generative AI.

On the contrary! With the +AI Pact + (which must precede a European regulation on AI, editor's note) that I proposed, we aim to help companies in their preparation, "he tweeted, accompanying his message with an image bearing the words "is this a threat?"

Microsoft boss Brad Smith went on to say that the drafters of the text would do well to listen to "those of us who may know more about the technical field" and who see "what is not likely to work".

Microsoft boss Brad Smith on November 12, 2022 in Paris © JOEL SAGET / AFP/Archives

"I have 35 years of experience working with the European Union (...) reason will prevail" because "legislation is a matter of compromise" and "AI is so important for the future competitiveness of companies," he said.

The project also worries Google, which sent its chief executive Sundar Pichai to Brussels this week. For now, faced with regulatory constraints, Google has given up deploying its Bard AI in the EU.

"Messiah"

Defending ChatGPT against overly restrictive regulations, while affirming that it must be supervised: this is the position of Sam Altman during his "OpenAI Tour", which takes him to sixteen metropolises.

On the program, tête-à-tête with heads of state - including Emmanuel Macron on Tuesday -, conferences and a hook to the Bilderberg group, the very discreet club of world leaders.

Sam Altman (r), boss of OpenAI and creator of ChatGPT, during a meeting with French President Emmanuel Macron at the Elysee Palace on May 23, 2023 in Paris © Yoan VALAT / POOL / AFP

This tour illustrates the stature of global AI guru acquired by Altman, 38, since the meteoric success of his chatbot, which has however triggered an avalanche of fears: misinformation, manipulation of elections, massive destruction of jobs and even threat to humanity.

A response from OpenAI is becoming urgent: many personalities have called for a pause in research, Italy suspended ChatGPT for three weeks for siphoning personal data and the G7 decided on Saturday to create a working group on AI.

The seduction campaign began on May 16 in front of US senators, with Sam Altman surprising by launching, in essence, "regulate me", and declaring that his nightmare was that AI would do "significant damage to the world".

Sam Altman, boss of OpenAI and creator of ChatGPT during a hearing before US senators, on May 16, 2023 in Washington © WIN MCNAMEE / Getty / AFP / Archives

But he also assured that many jobs would be created and stressed that too much regulation would hurt, because "if the American industry slows down, China or someone else can progress faster".

The next day, the young boss flew to Rio, then Lagos (Nigeria) and Lisbon, before Madrid, London, Paris, Warsaw, Munich and again Paris on Friday, at the Station F start-up campus. Next stops: Tel Aviv, Dubai, New Delhi, Singapore, Jakarta, Seoul, Tokyo and Melbourne.

"At Bilderberg, he was a little scary," said one participant, "but he also dangled the prospect of a country where to locate its European headquarters."

In Paris, Warsaw and Madrid, he was received with the respect of a head of state by political leaders eager to seize this economic opportunity, but who also recalled the need to supervise it.

Sam Altman's position during his "OpenAI Tour" is to defend ChatGPT against overly restrictive © regulations Stefani Reynolds / AFP/Archives

In Nigeria, he promised a flowering of start-ups and tried to restore the image of OpenAI, which used low-cost African workers to train his model.

In Rio, he hoped ChatGPT would generate "an improvement in people's living standards."

The reception was less consensual in London, where a handful of protesters demonstrated. "We shouldn't allow Silicon Valley multimillionaires with a Messiah complex to decide what we want," one student said.

© 2023 AFP