One of possible way that AI may evolve in the future may be that of a humanoid personal helper. These helper can work as personal housekeeper, baby sitter or even  have specialize coding that allow them to do more dangerous jobs like firefighter so that human don't have to risk there life for it. An example of this kind of AI robot was depicted in the the game: " Detroit Become Human" which take place in the far future where these type of helper or android are common place in workforce and household in the city. But just like in the game, we have to think about the possibility of AI coding being so advance that they are 99% human, having both logic and emotion. In this case, do we have to treat these android the same as a human, following human's laws and rights; as A whole new separate species that has their own laws and rights; or as just a plain old machine the same way as a computer we have now, you can modify them, sell them or abuse them as long as you owned them.

https://www.nbcbayarea.com/investigations/waymo-driverless-cars-safety-study/3740522/

Another can be that of a driverless vehicle that can act as transportation for us in the form of Waymo's robotaxi or as carrier for goods as the form of driverless container truck. In both cases, many proponent of AI say that this advancement will help decrease traffic accident as human error is not an issue; and the transportation itself is more efficient and faster as unlike human who need daily rests and breaks, an driverless vehicle doesn't need any of these. But there are some that think that not only is this taking away people livelihood, but also there is also some ethical issues with it. There is no way for an AI to learn every single type of accidents that may occur on the road as sometimes one just appear on the road and there will definitely be new type of obstacles that causes accident in the future. Let's say a driverless taxi with a passenger is in one of these accident, and the car is either going to have a big crash and kill the passenger or the car can swerve into a tree that only causes the passenger minor damage but will hit and kill a pedestrian near the tree. In this case how would the programmers code the AI and also who is responsible for the accident ? the programmers, the team leader or the company itself ?