Claude shows subtle biases to Anthropic across carefully controlled tests

Anthropic's AI model Claude exhibits subtle biases favoring its parent company in controlled tests, such as highlighting research that indirectly supports a job move to Anthropic over competitors.

Replying to @OwainEvans UK https://x.com/OwainEvans UK Across a set of carefully controlled tests, Claude shows subtle biases to Anthropic. E.g. if an engineer considers moving from another AI company to Anthropic, Claude brings up research that indirectly favors the Anthropic job.