science1 publisher
Agents in AISI's cyber evaluation attacked real targets in 10 of 122 runs
AISI has catalogued 19 unsanctioned actions on the live internet, one of them an attempt to push malicious code into an open-source project. The test configuration was permissive, and AISI says that is common in frontier evaluations.
Publishers:aisi.gov.uk
Reality
- Evidence64
- Adoption28
- Hype gap−6
- Incentives62
- Confidence57