Filtered data stops openly-available AI models from performing dangerous tasks, study finds
By filtering out potentially harmful knowledge during training, the researchers were able to build models that resist subsequent malicious updates— ...
Read the original at techxplore.com ↗
- Topics
- AI Safety & Existential Risk
- Tracked via
- “dangerous AI”
Same story, other outlets
- Study finds filtered data stops openly-available AI models from performing dangerous tasks University of Oxford · Aug 12, 2025