Yes, sure. AI is just like a child, it can learn. Depend on the supervised methods, they will get useful or harmful. When it has enough knowledge, they can learn by itself. This will be very similar with human.
Hi Andriana! It seems you are studying another field instead of AI, it is OK for researcher from other fields asking such question :) off course they can learn! the word is changing! now a days AI community trying to find some solutions like open box AI idea, to protect humanity. we are now in AGI (Artificial General Intelligence) state! (not like 30 years ago! writing AI program to play chess and put a lots of information on it about chess rule!) AGI can be like child! they don't know any information about the problem the should solve, (even the agent does not no whats problem!!!!) just do that by some approach like RL (Reinforcement Learning) to lean and solving problem without any prior knowledge! see famous nature 2015 paper By Deep Mind - forget about google deep mind :D (a very simple program learns every Atari 2600 games and masters on them buy just getting simple feedback from TV pixels) and some previous merit job by Juergen Schmidhuber teams
https://www.nature.com/articles/nature14236 Article Deep learning for real-time Atari game play using offline Mo...
they AI fields exponentially advanced per years, some prediction buy prominent scientific persons, guess that we will rich ASI (Artificial Singular Intelligence) near 2040-2050, which not only could lean but they can create things like humans!
Yes, sure. AI is just like a child, it can learn. Depend on the supervised methods, they will get useful or harmful. When it has enough knowledge, they can learn by itself. This will be very similar with human.