Agree with you -- it's valuable to validate assumptions if there is some controversy about those assumption. On the other hand, this work isn't even framed as

Agree with you -- it's valuable to validate assumptions if there is some controversy about those assumption. On the other hand, this work isn't even framed as a generalizable assumption that needed to be validated. It seems to me to be "just another example of how AI systems can be strategically deceptive for self-preservation."