Investing in Bengio

Today the Globe and Mail confirmed that Yoshua Bengio’s LawZero will be getting an investment of $300m from Canadian and German government sources. The purpose is to expand the efforts to build “safe” artificial intelligence, which Bengio is calling Scientist AI.

The new financing from Canada and Germany will be used to hire more staff and pay for the computational processing costs associated with research to build a system called Scientist AI that the organization says will not have the negative traits found in today’s models (Castaldo and Posadzki 2026).

How might things go wrong? According to Bengio, it comes down to three things:

First, the AI systems need to have the intention to do the bad thing – in the case of last summer, hack Hugging Face. Secondly, they need to be intelligent or capable enough to pull it off. And thirdly, the environment has to be sufficiently vulnerable for the attack to succeed. “All three were present last summer,” he said. (Castaldo and Posadzki 2026)

You can read more about the initiative on the LawZero website: https://www.lawzero.org

Importantly, ScientistAI will be trained without using reinforcement learning (RL). Reinforcement learning, according to Bengio, is one of the underlying reasons that we have misbehaving AI in the first place, and because it is embedded at the lowest levels of their training, it is difficult to remove (Bengio 2026).