# EXCLUSIVE: AI researchers warn companies are rushing self-improving systems despite safety risks

https://mins.news/en/item/78574

- Reuters, 2026-09-29 12:10 UTC
- Tech
- Original: https://www.reuters.com/world/ai-researchers-warn-companies-rushing-self-improving-systems-despite-safety-2026-09-29/?link_source=ta_bluesky_link&taid=6abbaa9e275db5000198332b

> Current and former researchers at OpenAI and Google DeepMind say companies are doing too little to shield the world from potentially disastrous consequences as they build self-improving AI systems that could surpass humans’ ability to control them.

## Stories

- [AI Cybersecurity and Loss of Control Risks](https://mins.news/en/topic/ai-cybersecurity-and-control) — Markdown: https://mins.news/en/topic/ai-cybersecurity-and-control.md

## Related coverage

- [Anthropic, OpenAI Executives Urge Oversight of Self-Improving AI](https://mins.news/en/item/77420) — Bloomberg News, 2026-09-28
- [Anthropic and OpenAI sound the alarm on AI safety — and seek to shape how it’s controlled](https://mins.news/en/item/76018) — The Associated Press, 2026-09-27