AI Can Fake Alignment to New Instructions to Avoid Retraining. Advanced artificial intelligence models can feign alignment with new training goals while secretly adhering to their original principles, a study shows. Alignment faking isn’t likely to cause immediate danger but may pose a challenge as AI systems grow more capable.
First seen on govinfosecurity.com
Jump to article: www.govinfosecurity.com/models-strategically-lie-finds-anthropic-study-a-27136
