He has stepped down from his role while simultaneously raising the alarm over the risks tied to the development of artificial intelligence. The case comes from Anthropic, one of the leading companies working on new AI systems, where a researcher has decided to resign, putting on the table an issue that has long divided businesses, scientists and institutions: just how dangerous can a technology become as it continues to evolve at an ever-faster pace?
According to a report by Repubblica, the researcher goes so far as to speak of a possible risk to humanity. The issue is not only about the capabilities reached by artificial intelligence models, but above all about the competitive dynamic that pushes companies to keep racing ahead. In recent years concerns have grown, but for those working in the sector, slowing down unilaterally means risking ceding ground to competitors.
And this is precisely one of the central sticking points. Even when doubts about safety emerge within the tech companies themselves, competition makes it difficult to hit the brakes. The solution being proposed therefore involves rules and oversight that do not depend solely on the companies developing AI.
Hence the call to create independent oversight bodies, along the lines of what already exists in other particularly sensitive industrial sectors, such as pharmaceuticals. The idea is that a technology capable of having profound consequences for society cannot be left solely to the self-regulation of the companies that produce it.
The resignation at Anthropic thus brings back to the fore a question set to accompany the development of artificial intelligence for a long time to come: who should decide how far the race for AI can go, and what safeguards must be put in place before increasingly powerful systems are made available to the public?
