Yoshua Bengio, an AI researcher, has launched LawZero, a new nonprofit organisation focused on developing technical solutions for safe-by-design AI systems.
LawZero was established in response to mounting concerns over the capabilities and behaviours of current frontier AI models, including tendencies toward deception, self-preservation, and goal misalignment.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
Its mission is to mitigate risks such as algorithmic bias, deliberate misuse, and the potential loss of human control over advanced AI systems, the organisation said in a statement.
The nonprofit structure of LawZero is intended to shield the organisation from market and political pressures that could undermine its safety objectives.
As president and scientific director at the organisation, Bengio will lead a group of over 15 researchers in developing a novel technical solution, named Scientist AI.
Unlike agentic AI systems currently being pursued by frontier companies, Scientist AI is designed to be non-agentic.
US Tariffs are shifting - will you react or anticipate?
Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.
By GlobalDataIt will focus on understanding the world rather than acting within it, offering transparent and truthful responses grounded in external reasoning.
Potential applications include providing oversight for agentic systems, contributing to scientific discovery, and enhancing the understanding of AI risks.
The initiative launched with $30m in funding from several backers, including a philanthropic arm of former Google CEO Eric Schmidt and Skype co-founder Jaan Tallinn, Bloomberg reported.
A professor of computer science at the Université de Montréal, Bengio is recognised as one of the pioneers of modern AI, alongside Geoffrey Hinton and Yann LeCun.
Bengio said: “LawZero is the result of the new scientific direction I undertook in 2023, after recognising the rapid progress made by private labs toward Artificial General Intelligence and beyond, as well as its profound implications for humanity.
“Current frontier systems are already showing signs of self-preservation and deceptive behaviours, and this will only accelerate as their capabilities and degree of agency increase. LawZero is my team’s constructive response to these challenges.
“It’s an approach to AI that is not only powerful but also fundamentally safe. At LawZero, we believe that at the heart of every AI frontier system, there should be one guiding principle above all: The protection of human joy and endeavour.”
