Yoshua Bengio joins ‘Safeguarded AI’ project

Turing Award laureate and AI pioneer, Yoshua Bengio, has been appointed as the scientific director of the ‘Safeguarded AI’ initiative. This project, funded by the UK’s Advanced Research and Invention Agency (ARIA), aims to enhance AI safety.

Writing in its bog, ARIA said: “We’re excited to welcome Professor Yoshua Bengio as scientific director for Safeguarded AI, supporting the work led by programme director ‘davidad’ Dalrymple. Yoshua’s work on the Cautious Scientist AI research agenda, which seeks to develop efficient methods for providing high-assurance guarantees that an AI system won’t take harmful actions, aligns closely with the Safeguarded AI programme.”

The Safeguarded AI programme, set to receive £59 million in government funding over the next four years, aspires to “usher in a new era for AI safety” by establishing quantitative safety guarantees through a “gatekeeper” AI system that monitors other AI agents.

“In doing so we’ll develop quantitative safety guarantees for AI in the way we have come to expect for nuclear power and passenger aviation,” the programme’s website states.

Bengio, a professor at the University of Montreal’s Department of Computer Science and Operations Research, shared the 2018 Turing Award with Geoffrey Hinton and Yann LeCun for their contributions to deep learning. He was also a signatory to a 2023 letter advocating for a pause in the development of large AI models until more robust auditing procedures are established.

An advocate for international collaboration on AI safety, Bengio chaired the International Scientific Report on the safety of advanced AI and attended the UK AI safety summit at Bletchley Park last year.

Bengio said that the complexity of advanced AI systems necessitates the use of AI to safeguard AI, as humans will soon be unable to comprehend the reasoning processes of large models.

A critic of voluntary agreements between governments and AI companies, Bengio prefers to place the burden of proof on these companies. He has proposed a democratic process for AI governance to ensure accountability and transparency, involving contributions from civil society representatives, similar to the regulation of the pharmaceutical industry.