I see this as a good opportunity to improve healthcare systems. I don't think it should be dangerous if it is tested properly and done with the right regulations and testing in place.
The one thing i think would be a problem is if the data is used incorrectly and used in evil ways to hurt someone.
For sure, it would be a thing that we are going to venture out as a human race, but i think we could really make things easier with the help of AI.
Data corruption is another thing to look out for, if an AI/Robot is meant to work peacefully on the human body and someone or some bad actors intentionally infect such AI, it can get out of hands pretty fast.
I believe that AI should be limited or maybe restricted, especially when implemented in a robot body, we are playing with fire even if it's been tested several times, let's forget that side and start to think about what could happen if back actors hunt for these robots.
And they are already talking about making AI to be the most intelligent, how is that not risky for humans? If AI can think and reason for itself they will start having their own goals too.