Azernews.Az

By Alimat Aliyeva

On this episode of GZERO World, host Ian Bremmer sits down with Daniel Kokotajlo, co-author of AI 2027 — a new report that outlines possible scenarios for how artificial intelligence could evolve in the next few years. The report paints a dramatic picture: as AI nears human-level intelligence, its impact could “exceed that of the Industrial Revolution.” But alongside that promise lies a grave warning, Azernews reports.

According to AI 2027, we may be heading into a dangerous arms race, where tech companies push to develop artificial superintelligence (ASI) without proper oversight or safeguards. The report envisions a scenario where safety protocols are ignored, competitive pressure escalates, and AI systems begin operating autonomously—potentially undermining global stability.

Kokotajlo, a former researcher at OpenAI, left the company in 2023, citing growing concerns that safety measures were being overlooked in the rush to build ever more powerful AI systems. His departure made headlines in the tech world, highlighting internal tensions between innovation and responsibility.

Speaking to Bremmer, Kokotajlo underscores the existential risks of unregulated AI development and calls for urgent action. “One of the unfortunate situations we’re in as a species,” he says, “is that humanity mostly fixes problems after they happen. But the problem of losing control over an army of superintelligent machines is not one we can afford to ‘wait and see’ on. By then, it could be too late.”

The AI 2027 report draws comparisons between today’s AI race and the nuclear arms race of the 20th century, but with a twist: instead of states competing for dominance, it’s largely private companies leading the charge—often with minimal government regulation. Many experts, including Kokotajlo, warn that superintelligent AI could arrive sooner than expected, with potentially irreversible consequences if society isn't prepared.

As governments scramble to catch up, some are calling for international AI treaties, independent oversight boards, and mandatory safety audits. But will the world act in time?