What changed
At the 2026 World Top Scientists Forum in Shanghai, Andre K. Geim and Martin Sweeting offered different answers to whether AI is dangerous. Geim called it a computational machine without perception or intent; Sweeting said it is unlikely to replace human thinking soon, but could challenge it over the long term.
Why it matters
The disagreement is useful because it separates two risks that are often bundled together. Geim rejects the image of a machine with its own will: AI, he said, has no “ghost” inside it and does not deliberately seek to harm people. The danger lies in how people use it, much as nuclear energy can generate electricity or be used to make an atomic bomb.
Sweeting’s concern is less cinematic and more practical. Today’s AI may be technically impressive, but it still needs human guidance and cannot match the breadth of human experience. Over time, however, its recursive development could put pressure on human thinking itself.
That makes the near-term question fairly clear: who directs the systems, and for what purpose? The longer-term question is harder. If AI continues to improve, it could challenge human reasoning in ways that are not yet possible to date or define.
The forum ran from October 30 to November 1 at Shanghai Lingang Center, drawing nearly 300 scientists and representatives of leading science-and-technology companies. Their exchange points to a technology revolution already changing how people communicate, learn, conduct research and develop industry, even if the more dramatic challenge to human thought remains a future possibility.
Comments
No comments yet.