Anthropic urges AI companies to slow down model development

14:29

The head of Anthropic, Dario Amodei, advocates for a more measured pace in enhancing the capabilities of artificial intelligence models. He proposes a three-step framework aimed at giving industry players more time to anticipate and address the risks associated with advancements in AI.

Dario Amodei warns against a too rapid technological race

Could the development of artificial intelligence systems be moving too fast? Dario Amodei, CEO of Anthropic, believes that companies in the sector should voluntarily slow down.

In an essay, the leader calls on AI stakeholders to moderate the pace at which they enhance their models' capabilities. For him, slowing down would not mean halting technological advancements but rather creating an additional timeframe to better manage their consequences.

According to Dario Amodei, even with a less vigorous pace, advancements in artificial intelligence would still be rapid enough to profoundly transform the industry.

Gaining time to better manage risks

The proposal from the head of Anthropic is based on the idea that time is an essential resource in the face of accelerating AI capabilities.

The goal would be to have a longer period to identify the risks that may accompany the evolution of models and to implement appropriate responses. This approach aims to align the pace of technological advancements with the measures needed to regulate them.

However, Dario Amodei emphasizes the need not to interpret this approach as a call to stop innovation. Instead, he advocates for a sufficiently controlled progression that allows companies, researchers, and public officials to adapt.

A debate at the heart of AI's future

Anthropic's stance comes in a sector marked by intense competition over the performance of artificial intelligence models.

Companies are continuously seeking to improve their systems, while discussions about the risks associated with these technologies are gaining more prominence. In this context, Dario Amodei's call brings to the forefront the question of the pace at which AI capabilities should progress.

His approach is based on a balance between innovation and caution: continue to move forward, but use the time gained from a more gradual development to better understand the consequences of each new technological step.