Andreas Kirsch voices concern about catastrophic AI risks
A DeepMind researcher expresses personal concern while qualifying what his warning says about current models.
I also am worried that AI will kill us all, either via near term risks or long term risks or both
About these dates
Statement date displayed on X in the inspected browser; not converted to UTC.
Google DeepMind research scientist Andreas Kirsch, who works on agentic language models, expressed fear that AI could kill humanity through near-term risks, longer-term risks, or both. Quoting Vishal Maini, he also described discomfort about making that concern public. Kirsch expresses concern without assigning a probability or fixed horizon. Public discomfort alone does not establish retaliation or the wording of current company speech rules. Andreas Kirsch ↗Andreas Kirsch ↗
His qualification distinguished current DeepMind models from the broader concern and rejected an immediate-doom interpretation. His post supplies a personal warning; Maini’s account of past communications practices remains separate testimony. Andreas Kirsch ↗Andreas Kirsch ↗
Sources & attribution
- Commentary 10 Sept 2026Personal warning about AI risks ↗
Andreas Kirsch. Original source inspected. Personal views.
- Commentary 10 Sept 2026Qualification about current models ↗
Andreas Kirsch. Original source inspected. Personal views.
- Organizational disclosure Accessed 10 Sept 2026Research biography ↗
Andreas Kirsch. Professional role checked; date is the access date.