AI Expert Warns About Need for Better AI Control

Yoshua Bengio, a leading figure in artificial intelligence and deep learning, has raised alarms about the urgent need for regulatory measures in the rapidly evolving AI landscape. In a recent interview, Bengio expressed concern that policymakers may be losing precious time to implement necessary safeguards as AI technology progresses at an unprecedented rate. His perspective shifted dramatically following the emergence of OpenAI's ChatGPT in late 2022, which led him to reassess his previous belief that human-level AI was decades away. Bengio's advocacy for AI regulation comes at a time when many in the tech industry, including venture capitalists and academics, oppose legislative efforts like a proposed California bill, fearing it could hinder innovation and U.S. competitiveness. He argues that the discourse surrounding AI regulation is often influenced by those seeking to profit from the technology, emphasizing the need for swift action. "Policy takes a lot of time," he noted, warning that there may not be a decade to address these challenges effectively. While acknowledging the contentious nature of AI's potential risks, Bengio insists on the importance of protecting the public from worst-case scenarios, similar to existing government regulations in other sectors. He highlighted alarming applications of AI, such as synthetic child pornography and deepfakes, which could undermine democratic processes. Despite these concerns, Bengio remains hopeful about AI's benefits in areas like biology and medicine, advocating for significant investment in these safer applications.