# Outgoing OpenAI safety engineer warns AI giants are ‘not being nearly careful enough’

https://mins.news/en/item/84921

- DW, 2026-10-03 21:55 UTC
- Tech
- Original: https://www.dw.com/en/ai-giants-not-being-nearly-careful-enough-warns-outgoing-openai-safety-engineer/a-79533080?maca=en-rss-en-all-1573-rdf

> Former OpenAI safety engineer David Robinson says AI companies need safeguards comparable to those in nuclear power and aviation. He argues that safety research must come before more advanced models, while OpenAI says it pauses training or withholds models when they cannot be safely managed.

## 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

- [OpenAI safety employee quits, says ‘time for trial and error is over’](https://mins.news/en/item/85866) — CGTN, 2026-10-05
- [Robust AI co-founder Gary Marcus warns of inadequate safeguards for uncontrollable AI systems](https://mins.news/en/video/85719) — Bloomberg Television, 2026-10-04
- [AI needs safety layers like aviation or nuclear plants, says ex-OpenAI engineer David Robinson](https://mins.news/en/item/84914) — TRT World, 2026-10-03