Much effort went into creating sims, uploaded personalities obtained through destructive scanning of a person's brain. These efforts have largely been successful.
Whilst officially banned in the civilized world, the practice has gained significant currency, and terminally ill rich people now routinely travel to Transnistria, where you can still buy bullets by the bucket, semi-sentient anti-aircraft missiles, and yes, get a (sometimes dubious quality) mind upload done with no legal hassles.
The Sims are passing the Turing test with no issues, and for the most part (save for the occasional abomination), sound indistinguishable to living friends and relatives from the original. Increasingly, the sims are fighting to gain legal recognition as "Remnanters" - legal successors of the dead Person they used to be. However, problems soon crop up in dealing with identity. One can (and sometimes does) create multiple copies of the same uploaded personality. So Richie Rich travels to Transnistria, gets their brain "read" and creates a Richie Remnanter A, who is placed in an android body and walks the Earth again. A different version of Richie Rich, Richie Remnanter B, chooses to stay in a virtual environment running at 10x real-life speed.
I have thought of 3 approaches:
Remnanter A and Remnanter B are now different persons and diverge exponentially from here on. This is the default path, but leaves thorny legal issues, such as which Remnanter gets control of what portion of assets.
Somehow, only one Remnanter is allowed. This would be hard to implement technologically, especially with ruthless Romanian and Russian hackers and bio-engineers in charge.
Remnanter A and Remnanter B occasionally sync up, creating a unified personality with multiple presences. This seems hugely technologically challenging, even more so if we want a (near-)continuous sync.
Which one of these (if any) is most likely to be the preferred path, and why? If you think another outcome is more likely, I would like to hear it instead.
Why Sims? It turns out that creating non-insane AIs is hard. In retrospect, AI research historians suggest, it should have been obvious that the possible Hilbert space of insane dysfunctional minds is much, much larger than the space of sane minds, and without the finely tuned evolutionary help of millions of subjective years of evolution, one is highly unlikely to stumble into the sane space. Programming AI from scratch has thus proven impossible.