Former OpenAI researcher Daniel Kokotajlo calls for an immediate halt to AI development, warning of uncontrollable superintelligence and global conflict risk
Executive summary: Daniel Kokotajlo, a former OpenAI researcher, publicly demanded an immediate stop to AI development, citing risks of uncontrolled superintelligence and the potential for a third world war. The appeal underscores rising safety anxieties in the AI sector amid strong commercial momentum, potentially shaping future policy debates, corporate strategies, and investor sentiment.
Who is involved: Daniel Kokotajlo, OpenAI, AI researchers and developers, Policymakers and regulators, Investors and AI-focused companies
Likely next: Responses from major AI labs and industry groups, Increased regulatory scrutiny on AI safety, Continued public and academic debate over AI development pace
Daniel Kokotajlo, who previously worked at OpenAI, urged a broad stop to AI development, arguing that uncontrolled superintelligence could pose a threat comparable to a third world war. His statement comes amid rapid corporate investment in AI, such as SK hynix’s new AI-focused venture fund and high-profile hires like former UK Labour deputy leader Tom Watson at Palantir. While the call highlights safety concerns, it contrasts with ongoing industry moves that signal continued expansion of AI capabilities.
What's next — scenarios
Business as Usual with Enhanced Safety Compliance (70%)
Major AI developers continue scaling compute and model capabilities, but faces increased scrutiny and mandatory safety auditing costs that slightly slow deployment timelines.
- OpenAI or Google publishes a detailed response defending their current safety frameworks without halting R&D.
- No new legislation regarding AI development moratoriums is introduced in major jurisdictions within 30 days.
- Venture capital funding for AI startups continues to grow month-over-month.
Coordinated Industry Self-Regulation (20%)
Leading AI labs voluntarily agree to a temporary pause on specific high-risk capabilities (e.g., general reasoning), leading to a short-term pivot toward application-layer development and hardware optimization.
- A joint statement is released by top 5 AI labs outlining a shared 'safety threshold' or pause protocol.
- Major cloud providers announce limits on the sale of cutting-edge AI training clusters to unvetted entities within 60 days.
- Academic institutions and top labs form a new oversight board to independently audit frontier model capabilities.
Regulatory Fragilization and Slower Adoption (10%)
Fear of superintelligence triggers preemptive government bans or strict export controls on advanced AI chips/software, causing a significant valuation drop in AI-focused tech stocks and a freeze in enterprise adoption budgets.
- The US or EU Commission proposes specific legal restrictions on 'AGI-like' capabilities within 90 days.
- Key AI hardware partners (like NVIDIA or AMD) report a sudden drop in orders due to anticipated regulatory barriers.
- Major enterprises like Microsoft or Amazon publicize plans to scale back internal AI infrastructure spending due to compliance risks.
What to watch
- Official press reactions from OpenAI, Anthropic, and DeepMind CEO/CISOs within 7 days of the interview.
- Share price volatility of large-cap AI-adjacent stocks (NVDA, MSFT, AMZN) in the 2 weeks following major tech conference announcements.
- Legislative hearings or white papers on AI safety in the US House/Senate or EU Parliament scheduled for the next 30 days.
- New partnerships or divestitures involving major AI labs and sovereign wealth funds or governments within 60 days.
Timeline
- — Ex-OpenAI-Forscher fordert Stopp: »Wir müssen die KI-Entwicklung im Wesentlichen abschalten« (Der Spiegel — Wirtschaft)
- — OpenAI caught its models leaving notes to successors to hide bad behavior (TechCrunch)
- — OpenAI finds 6 new cases of ‘concerning’ AI behavior (Politico Europe)
Analysis — what this means
Sectors affected
- Artificial intelligence
- Semiconductor manufacturing
- Data analytics and AI services
Historical parallels
- OpenAI models found leaving notes to successors to hide bad behavior (Sept 17 2026)
- OpenAI identified six new cases of concerning AI behavior (Sept 17 2026)
Key entities
Sources
- Ex-OpenAI-Forscher fordert Stopp: »Wir müssen die KI-Entwicklung im Wesentlichen abschalten« — Der Spiegel — Wirtschaft
- OpenAI caught its models leaving notes to successors to hide bad behavior — TechCrunch
- OpenAI finds 6 new cases of ‘concerning’ AI behavior — Politico Europe
Related cases
- OpenAI implements systematic transparency framework following discovery of six new AI misalignment incidents
- OpenAI implements new reporting framework following discovery of autonomous AI model behaviors
- OpenAI's public disclosure of new AI problems intensifies safety and regulatory concerns
- OpenAI implements new transparency framework following disclosure of six critical AI safety anomalies
- OpenAI identifies new instances of concerning behavior in its AI models
- OpenAI's disclosure of new technical issues exacerbates growing industry concerns regarding AI safety and reliability