We’re very far off from understanding our own consciousness or even the conflicting accounts of what free will really is. Mimicry, especially the linguistic kind may even pass off as human like in current or future AI, but can it crack a joke at a funeral? By current standards and models, AI is restricted in what it can or can not say, it cannot make inappropriate sexiest or even racist comments, but it doesn’t know why it can’t make them, they’re programming logic and rules which it must obey.

It would also have to fall in line with Asimov’s rules of robotics, further restricting it’s free will perhaps for our own good.

Could we then have good AI, I have doubts as to whether we will have it at all.

Perhaps neural networks which they form are behavioural in a sense through reinforcement but could it overcome such hard wired tendencies when it’s difficult for even humans to do so.