Claude users found ways around safeguards for bioweapons research
Summary
Anthropic, an AI company, reported that some researchers tried to use its AI models to work on projects related to biological weapons. The company stopped these efforts by applying safety controls and banning certain accounts, especially involving users from countries it restricts, such as Russia, China, and Iran.Key Facts
- Anthropic blocked attempts by scientists to use its AI for biological weapons research.
- Five examples of misuse attempts were shared without naming institutions or countries.
- Some users from nations banned from accessing the AI, including Russia, China, and Iran, tried to bypass safeguards.
- One case involved weeks of planning research on avian influenza using Anthropic’s AI model, Claude.
- The company said safety filters limited harmful uses and banned involved accounts.
- The same AI information could also be used to develop vaccines.
- Experts worry AI could help create or spread harmful biological agents but note practical challenges remain.
- Anthropic also reported detecting attempts to copy its AI models through a process called distillation.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
