Fei-Fei Li, a professor at Stanford University and co-founder of World Labs, has called for increased oversight of artificial intelligence by independent bodies and the public sector. She argues that relying solely on internal assessments by AI companies is insufficient for evaluating the safety and capabilities of increasingly powerful systems.
Li emphasized that while internal research is valuable, it cannot replace broader standards established through collaboration between governments, industry, and academic institutions. She advocates for a balanced approach where independent organizations play a key role in setting benchmarks for AI safety and performance.
The debate over AI risks has intensified among industry leaders. Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have both recently suggested slowing the pace of frontier model development. These discussions are set to continue at the United Nations Security Council, where safety concerns remain a primary focus.
Li noted that the challenges of managing AI are comparable to historical technological shifts, such as the introduction of electricity or steam engines. She stressed that society must maintain control over technology to ensure it aligns with human dignity and autonomy, rather than focusing solely on the technology itself.
Meanwhile, the political landscape remains divided. During the UN General Assembly, former US President Donald Trump dismissed concerns regarding existential threats from AI, labeling them a hoax and advocating for the continued acceleration of technological development.
Li is currently focused on developing 'world models' at her company, World Labs, which generate and reconstruct 3D environments. She highlighted that as these systems transition into real-world applications like robotics, design, and healthcare, establishing robust evaluation frameworks becomes a critical industry priority.
