Ever since ChatGpt and other AI have become popular. There has been talk about how far this technology can go. There are even some true believers in AI who believe that this tech can reach the Artificial General Intelligence stage. In that the machine itself has reached an intelligence matching humans, it can make decisions on its own. This is the ultimate goal of the people working on AI. On the other side, this is what scares a lot of people about AI. If AI becomes intelligent they wonder about all sorts of scenarios which are written in books and shown in movies, where robots take over the world!
The debate that interests me is how do we test so we know AGI has been achieved? If passed we can say that AI has reached that stage. In one podcast a person suggested that, if a new movie is released the AI ‘sees’ the movie, but does not get to see any review or read up about it(so no beforehand knowledge). If it gives out a good analysis/review of the movie then it can be said AGI has been reached. It’s a good test, so we just need to think of ways to test the AI in a way we test humans, humans have a way of figuring things out, if AI can figure things out on its own without help(like pulling existing data, producing answer from some other source) that we could say it has reached AGI.
I do not believe true AGI can be achieved. It could probably reach a point that is close to it, that though is still a long way off, but to be like human intelligence I don’t think is possible. The complexity of the way we think, and how our mind works, even with so much science behind it, is still not fully understood well. Humans are just special, for machines to reach that is nearly impossible. Saying that it looks like the advancement of AI will become super useful to humans, and will be a major advantage for people who use it well, against people who have less knowledge about it. AI is likely going to be very disruptive in every field it reaches.