DaveW ... interesting avenue ... I guess it depends just how 'human' they become ... you might imagine that, as a coldly intellectual 'being', it would have no reason to be anything other than entirely truthful, as it wouldn't have the ego imperatives to.protect (and the overwhelming number of Human lies are minor and to protect our self-view/social-standing) ... it wouldn't give a toss about that. However if it became Human-like it might also pick that baggage up along the way.
More worrying would be an AI that had been 'brought up' with a deliberately skewed input or with it's access to information rigorously controlled (think Einstein raised by a radical cult.) It would follow the only way it knew.
Similarly, without empathy or a moral framework with a broad knowledge base then lies would be seen as just another tool to acheive its alloted task and positively useful..
My worry is powerful synthetic intellects moulded by imperfect Humans at best ... and deliberate bad-actors at worst.
K
