“It will be difficult for Europe to compete at the AI border”

“ Ai believe, in fact, that it will be very difficult for Europe – or for any other country other than the USA or China – to actually compete at the border, simply because of the high cost and the challenges involved, as well as the speed with which [AI] technology evolves,” says Malo Bourgon, who spoke to Lusa from Berkeley, USA.

While considering that there is room for all the different world powers to start talking about artificial intelligence (IA), obviously “the two main protagonists in the development of this technology are currently the US and China,” says the CEO of the non-profit research institute based in Berkeley.

“ The world is divided in various ways and there are many trust issues between the US and China, to say the least, but things can also change very quickly,” he admits.

In fact, with the launch of the Mythos [Anthropic] “we went from a scenario where the US government was mainly talking about letting the AI industry take its own course — while seeking, in a way, supremacy in the area — to a situation where, very quickly” the US imposed export controls on the most powerful model in existence, reaching the point of not even allowing the Americans themselves to use it effectively, he recalls.

On China’s side, “it’s much harder to come to an understanding, but they are carrying out various internal regulation actions, and their public statements have been much more cautious about issues such as AGI [General Artificial Intelligence or General Artificial Intelligence] and superintelligence, as well as what that might mean,” he continues.

In the framework of Trump’s trip to China, the desire was expressed to have dialogue with the Chinese on the safeguards for the AI, so “they are all initial signs”, but “they are not solutions to the problem.”.

Malo Bourgon points out that in the last six months the conversation about AI has changed as the capabilities of technology have awakened to how powerful models can be.

So it will not be easy, “but it seems possible” for governments to start discussing the development of the AI.

Malo Bourgon admits that it is “very difficult for many countries and companies to compete”, given the requirements of talent and high investment in infrastructure needed to build state-of-the-art AI systems.

Therefore, “I believe that — with the exception of major US companies (such as Google, Anthropic and OpenAI), which lead technological advancement in this area, and some Chinese companies that follow them closely — it will be extremely difficult for most other actors in the sector to compete with these companies, unless they invest much higher amounts and attract the best talents to a much higher scale,” he stresses, which removes Europe from the equation.

Malo Bourgon says he believes that “the most powerful role” that Europe can play is to “express its concerns on the international scene about the course of this technology and its course”.

About the AI, the person responsible says it is “a kind of metatechnology”.

That is, “it not only develops faster and has the potential to be more disruptive than any technology of the past, but it is a technology that will help us — as it becomes more able to act on its own — to invent new and even more powerful technologies,” he points out.

“ Despite the risk that worries me — that of the loss of control and extinction of humanity — and even disregarding the risks of national security associated with the dual use of models with increasingly powerful capabilities, these are the objectives of these companies”: to build AI systems capable of essentially carrying out “that type of cognitive activity that we perform”, similar to that of human beings.

In a sense, “it is about automating the automation process. And I believe that it makes sense to think of the future versions of this technology increasingly as another intelligent species that we are introducing on the planet, rather than just tools that humans use in the course of their work or the activities they carry out,” he says.

AI companies “are talking about how they believe there is a 60% chance that by the end of 2028 they will automate the very process of research and development” of technology, and the AI systems will build their successors, “which will be more powerful and smarter, more capable of creating the next generation.”.

Therefore, “I hope this is the most powerful and disruptive technology humans have ever developed in their entire history,” Malo Bourgon concludes.