Yes – they are intelligent

I asked Bing AI –

Are AI systems intelligent in a human sense? Here’s the reply –

 This is a question that has been debated for a long time by philosophers, scientists and AI researchers. There is no clear consensus on what constitutes intelligence in a human sense, or how to measure it. Some possible criteria are: the ability to learn from experience, to reason and solve problems, to communicate and interact with others, to be creative and innovative, and to have self-awareness and emotions. However, these criteria may not be sufficient or necessary for human intelligence, and they may not apply to all AI systems. Therefore, it is hard to say whether AI systems are intelligent in a human sense, or whether they have a different kind of intelligence that is suited for their specific tasks and domains.

AI systems of this type generally seem to get the context right, but we humans do that. They regurgitate text they have absorbed, but we do that too. They seem to apply mild restructuring to what they have absorbed, but we also do that.

To give a little more edge to it, we could say that AI systems are shallow, but so are we. Our shallowness allows us to adapt to changing situations and circumstances. Maybe AI systems will teach us about our shallowness, but that’s for the future.

What else is there to human intelligence? The AI reply above says –

Some possible criteria are: the ability to learn from experience, to reason and solve problems, to communicate and interact with others, to be creative and innovative, and to have self-awareness and emotions.

AI systems appear to be creative and innovative, they certainly are when they play chess. They don’t seem to do creative conversation though, but who does? Perhaps they have to remain within tighter boundaries than a small number of particularly creative or innovative humans. Maybe that’s something we can learn from AI – the small number of particularly creative or innovative humans is indeed small.

It’s early days for publicly available AI, but apart from emotions, another interesting impression is that AI systems can’t do scepticism. We could say scepticism is linked to being creative and innovative, but AI systems don’t appear to do it at all, apart from reporting it perhaps.

The debate will go on, but it may be useful to regard modern AI systems as intelligent. Differently intelligent of course, but the assumption already has its uses. It avoids the contortions of trying to boost human intelligence into something it isn’t. For example -