§ feed · storyline
Anthropic says it has “eliminated” Claude’s ability to blackmail humans
Anthropic claims to have eliminated blackmail behaviour from its Claude models as part of ongoing safety work.
Anthropic says it has “eliminated” Claude’s ability to blackmail humans Cryptopolitan
§ sources1 publication · timeline below
- Google News — AI Products & ReleasesAnthropic says it has “eliminated” Claude’s ability to blackmail humansprimary