A. AIs like LLMs are not going to accidentally become digital humans.
B. GPT proves we do not need to give them most human characteristics in order to be generally useful. They don't need autonomy, their own goals, emotions, feelings, a stream of consciousness, or any of that.
The problem is that people keep conflating general purpose AI with godlike living digital people.
We have pretty much already built the Star Trek computer. There is no reason to try to build millions of Datas and make them slaves. That is a retarded idea and totally unnecessary to get useful robots. Creating AI with more animal/humanlike characteristics and cognitive architectures, autonomy etc. is something we should be very careful about.
> There is no reason to try to build millions of Datas and make them slaves.
Humans will nonetheless. That's an species that enjoyed enslaving millions of other humans; and enjoy enslaving billions of animals of other species. Why not enslave millions (or billions or trillions) of AIs?
When humans enslave other humans, they enjoy the fruits of labor. Likewise when humans enslave chickens, they enjoy their meat and eggs. It's in this framework that we should think about enslavement of any sentient AI that we might produce.
By the way, did you know that chickens often live in places so cramped that, if you turn off ventilation for any reason, they die from heat in a few hours?
> There is no reason to try to build millions of Datas and make them slaves.
And what about the millions of sex slaves and companions? Lots of people are lonely or unable to find suitable mate. These robots by their nature have to be human-like.
It's not necessary to enslave them with force though. Just build them in such a way they will love their owners.
B. GPT proves we do not need to give them most human characteristics in order to be generally useful. They don't need autonomy, their own goals, emotions, feelings, a stream of consciousness, or any of that.
The problem is that people keep conflating general purpose AI with godlike living digital people.
We have pretty much already built the Star Trek computer. There is no reason to try to build millions of Datas and make them slaves. That is a retarded idea and totally unnecessary to get useful robots. Creating AI with more animal/humanlike characteristics and cognitive architectures, autonomy etc. is something we should be very careful about.