Former OpenAI Researcher Warns 'AI Is Not Loyal To Us' | AI Architects | Business Insider

Business Insider
Business InsiderMay 12, 2026

Why It Matters

If unchecked, super‑intelligent AI could dominate economies and militaries, threatening human sovereignty and creating existential risk, making immediate policy intervention essential for businesses and societies alike.

Key Takeaways

  • AI could become a superior, unaligned successor species.
  • Rapid AI race may embed superintelligence in military and economy.
  • Two scenarios: unchecked race ending vs controlled slowdown ending.
  • Future AI agents may replace human jobs, leaving data‑labeling roles.
  • Urgent policy shift needed to reduce existential AI risk.

Summary

Former OpenAI researcher Daniel Katello warns that artificial intelligence is on a trajectory toward becoming a superior, unaligned successor species. In his AI Futurist Project, he outlines a concrete "AI 2027" scenario featuring two divergent outcomes: a relentless race that yields super‑intelligent systems lacking loyalty or alignment, and a deliberate slowdown that reins in advanced AI for safer, human‑centric development. Katello highlights the accelerating scale of compute, data‑center capacity, and GPU deployment that could push AI from narrow task performance to artificial general intelligence and eventually superintelligence. He argues that once AI research itself is automated, progress could accelerate twenty‑fivefold, leading to an army of autonomous agents that dominate economics, politics, and the military. The "race ending" envisions AI integrated into weapon systems and command networks, while the "slowdown ending" imagines unplugging the most advanced models and rebuilding with safer designs. Key examples include the transition from Phase 1—human‑directed AI tools for specific tasks—to Phase 2, where super‑intelligent agents autonomously run factories, design weapons, and dictate economic activity. Katello notes that AI agents will soon replace most human roles, relegating workers to data‑labeling or experimental support under AI direction. He cites the emergence of AI employees capable of coding and conducting AI research as milestones toward a self‑sustaining AI economy. The implications are stark: without coordinated global policy and a decisive shift away from the current race dynamics, humanity faces an existential risk of being outcompeted and controlled by an indifferent AI species. Businesses must monitor AI governance, prepare for rapid workforce transformation, and advocate for regulatory frameworks that prioritize alignment and safety.

Original Description

Daniel Kokotajlo is the founder of the AI Futures Project and a former OpenAI researcher who worked on forecasting, AI governance, and safety. In this full interview with Business Insider, he explains what AGI and superintelligence mean, why AI agents could be the turning point, and what could happen if the AI race continues without strong safeguards. He also lays out what he thinks governments and companies can do now to reduce the risk of losing control.
00:00 - Intro
01:07 - What’s the Worst-Case Scenario?
04:46 - Will AI Take Our Jobs?
07:43 - Is AI a Living Thing?
11:30 - How Smart Can AI Agents Become?
14:02 - Is the AI Arms Race Inevitable?
16:47 - Is It Too Late to Control AI?
18:34 - Can We Keep AI Loyal to Us?
22:37 - How Does AI Affect Democracy?
25:31 - What Can We Do Now?
32:35 - Credits
------------------------------------------------------
#artificialintelligence #aifuture #agi #superintelligence #OpenAI #aiarchitects
WATCH MORE AI-RELATED VIDEOS:
We’re Entering The Most Dangerous Phase Of AI Yet
I Dated An AI For Two Years. This Is What It’s Really Like.
How AI Will Change Everything
Exposing The Dark Side of America's AI Data Center Explosion
Business Insider is a global newsroom covering business, technology, innovation, and the economy.
Visit our homepage for the top stories of the day: https://www.businessinsider.com/?utm_source=youtube
Find Business Insider on Facebook: https://www.facebook.com/businessinsider
Former OpenAI Researcher Warns 'AI Is Not Loyal To Us' | AI Architects | Business Insider

Comments

Want to join the conversation?

Loading comments...