Is another artificial intelligence possible? In Switzerland, the model of responsible tech

Is another artificial intelligence possible? In Switzerland, we find a model of this responsible tech. In the suburbs of Geneva, just a stone's throw from the French border, the basement of an eco-neighborhood houses the Infomaniak servers. It is the largest digital host in the country.

Created in 1994, this company of 200 employees was built as a counter-model to the heavyweights of the sector, such as the Americans Meta, Google, OpenAI or Anthropic. It also belongs to a public utility foundation. The latest addition to its services is called Euria. “It’s Infomaniak’s AI assistant. When you get the application on your mobile phone, you find yourself with a chat. Exactly the same way you can have with Claude or ChatGPT,” explains Thomas Jacobsen, Infomaniak spokesperson.

Except that Euria is a discrete AI. It does not share or resell user data, that is to say all the information that we entrust to the AI. Infomaniak has its own data centers, all in Switzerland. And to guarantee this data security, it takes a digital version of one of the country's great specialties. “The data that is entrusted to Euria is like a safe that cannot be exploited from the outside. The software layer that is developed by Infomaniak in Switzerland does not depend on the outside to function,” he explains. Also, the heat recovery system produced by the data centers is then used to heat 6,000 homes in order to limit its environmental impact.

Can this type of initiative influence the sector? The question makes Sihem Amer-Yahia, research director at the CNRS in Grenoble, smile. She knows how Big Tech works. “It’s organized, but under other skies than ethics,” she laughs. It’s almost in the DNA of pioneering companies, she says. "The web has not existed for that long but there is a history which has established that these companies operate by jurisprudence. Technologies, companies have advanced so quickly that there has been no legislation to regulate the use of platforms. And therefore, legislation is done by reaction", analyzes the researcher.

The big bosses of Silicon Valley are calling for regulation of the sector. They even fear that AI will bring about the end of humanity. But for Jan Fergusson, sociologist at INRIA, the Digital Sciences Research Institute, these calls are not really synonymous with ethical awakening. “Existential risk serves as a bit of a counter-fire to much more real subjects such as the ecological footprint and the way in which certain workers are exploited. AI leaders never talk about all of this,” he denounces. Juggernauts like OpenAI, the father of ChatGPT, account for 95% of AI traffic. On the other hand, the Euria AI is only aimed at French speakers worried about their data. In the world of AI, ethics is still a very small niche.