The AI Can Now Create a Replica of Your Personality article in MIT Technology Review talks about how the researchers can create simulation AI agents which behave and make decisions as a specific individual would. The agent design is based on a two hour interview. One possible use for these agents is doing social science research on human behavior where the cost and risks for human subjects is too high.

There are three examples of real world AI applications in this article that are being used by companies for decision making. They are not personality based like the example above but are expertise based, another example of human like decisions with AI.

On the other side of the spectrum is the perspective from an HBR article that Generative AI is Still Just a Prediction Machine reminding us that AI agents are not human or even close to human and applications need human judgement of when and how to use.

All of this raises lots of ethical concerns and this article from MIT Technology Review talks about the ethical concerns of simulations agents like the above. Two related and specific concerns are the impact this type of AI has on creating more realistic deep fakes and whether we have the right to know that we are talking to an AI agent versus the real human.