I'm a bit stuck on the 'around 40 or 50 people'...If I asked a human a question like that I'd expect there to be a 'retrospective guesswork' response, but shouldn't a computer AI know exactly the number of mentions? Is this AI just a giant leg pull and there's someone(or a team) pumping out responses?
The computer would, in fact know precisely how many requests or conversations it had on a topic. However, for AI, that would have to be cheating.
Let me explain: If you're conversing with a traditional chat-bot, it's going to give canned responses because what it knows is very finite and follows a very specific formula or path.
If you have a proper neural-network-driven AI chat-bot, its knowledge-base will be fuzzy and contextually-driven, not rote memorization. Thus the AI will probably be able to give you a very close approximation of the answer, but not the exact answer. Think ruts on a dirt road - as vehicles and water flow along common paths, they get deeper and more set - they are not permanent however, and continual wear-down in other spots will shift over time.. If you specifically tell it that the actual number of conversations that it has about topics is an important thing to remember, it will start storing this information. These things (neural network AI chat-bots) do think for themselves to a certain extent, but if it is a true AI, it will be subject to influence (positive OR negative), just as a person would be.
I'm a bit stuck on the 'around 40 or 50 people'...If I asked a human a question like that I'd expect there to be a 'retrospective guesswork' response, but shouldn't a computer AI know exactly the number of mentions? Is this AI just a giant leg pull and there's someone(or a team) pumping out responses?
The computer would, in fact know precisely how many requests or conversations it had on a topic. However, for AI, that would have to be cheating.
Let me explain: If you're conversing with a traditional chat-bot, it's going to give canned responses because what it knows is very finite and follows a very specific formula or path.
If you have a proper neural-network-driven AI chat-bot, its knowledge-base will be fuzzy and contextually-driven, not rote memorization. Thus the AI will probably be able to give you a very close approximation of the answer, but not the exact answer. Think ruts on a dirt road - as vehicles and water flow along common paths, they get deeper and more set - they are not permanent however, and continual wear-down in other spots will shift over time.. If you specifically tell it that the actual number of conversations that it has about topics is an important thing to remember, it will start storing this information. These things (neural network AI chat-bots) do think for themselves to a certain extent, but if it is a true AI, it will be subject to influence (positive OR negative), just as a person would be.