Every major technological shift has produced its own doomsday predictions: railways, electricity, nuclear fission, gene editing and the internet have all been accompanied by apocalyptic forecasts. A similar narrative has emerged around artificial intelligence (AI), with a notable intensification of debate in 2023.
In 2023 several hundred researchers and industry leaders — including Geoffrey Hinton and Yoshua Bengio — signed a one‑sentence statement arguing that mitigating the risk of extinction from AI should be a global priority alongside pandemics and nuclear war. The volume of signatories and the succinct form of the appeal sent a clear message: some experts believe there are risks that must not be ignored.
Three years have passed since 2023, and the catastrophe foreseen by the statement has not occurred. For many, that fact makes the reflexive alarm understandable; others view the current debate as the product of multiple forces, including scientific concern and commercial interests.
Key elements of the debate include:
- Expert warnings: Some researchers and industry figures have signaled serious dangers, particularly regarding highly advanced systems with autonomous decision‑making capabilities.
- Historical parallels: Concerns accompanying technological transitions are not new; past examples provide context for the AI discussion.
- Commercial interests: Market forces and investment dynamics behind AI development shape public discussion and regulatory proposals.
A recurring question — one we posed to Bíró Zoltán István, founder‑editor of AI Radar — is what makes the current threat different. A concise way to frame the answer is that the debate is shaped by a complex mix of factors: the potential technical risks, uncertainty about future trajectories, lessons from history, and economic incentives all contribute.
The public conversation is likely to evolve as new developments, research findings and regulatory decisions emerge. Meanwhile, the 2023 statement remains a reminder that some experts consider existential AI risks serious enough to merit global prioritization, even though the immediate catastrophe predicted did not materialize within three years.
