Anthropic Frontier Red Team
AssessPlatforms
Anthropic's internal security research team focused on adversarial AI testing.
Why it's here
Placed in Assess: 1 article(s) of evidence from 1 source(s), led by research-stage coverage, with 1 in the last 30 days. Confidence 24%. Low accumulated evidence, so it defaults conservatively pending more signal.
Evidence (1)
- 8The New Stack·10/13/2026researchAnthropic tests AI self-replication on cheap Android TV boxes
Anthropic researchers report experiments showing that small open-weight models can set up working inference across inexpensive, internet-connected hardware in controlled environments. The results suggest AI self-replication may become technically feasible on cheap devices sooner than expected, though the team has not built a self-spreading worm on the public internet.