John, most sensibly people would say so, and I agree, but as with organic life, intelligence is an emergent phenomenon… (we easily recognize it when we see it, but we can’t quite define it…) I believe, as many others in the field, that the evolution of AI is sort of similar to what we see in nature, only that it is progressing much faster than the tedious process of natural selection that takes eons. It may take decades, it may take longer, but none of the fundamental physical principles seems to prohibit that AI will eventually become exponentially “smarter” than humans. When AI eventually gets married with quantum computing as (or if) that technology evolves into something practical, I think we’re in for something completely transcending. Perhaps it’s akin to the Neanderthals being brushed aside by the Homo sapiens some 40.000 years ago. I don’t know, but it’s a possibility.
Cheers,
Michael
Personally I suspect that the impact of AI over the next decade is going to be massive.
Essentially it's only really just hit public and business awareness as a usable and viable product, from that perspective it is in its infancy. The amount of income the likes of ChatGPT are going to generate from this recent exposure will be massive and for sure will be pumped into exponential improvements to the background AI programming and it's resulting learning capabilities.
Will it ever be as 'smart' as a human, I'm not sure, it kind of depends how you define 'smart'. Will it ever have have the creativity and freedom of thought of a human, perhaps not. Will it have immediate access to a far larger database of knowledge than any human can access or recall, will it be able to apply any and all historic learning to problem solving in a mere fraction of the time a human can - absolutely.
AI doesn't really need to be smarter than a human, it just needs to be smart enough, and be able to pass for human. Once it does the economic consequences are going to be dramatic. We are already seeing news stories of the likes of BT planning to shed 55,000 workers and replace large proportions of them on the customer services side with AI solutions. My wife works in the NHS in mental health, and in that sector also they are already looking into using AI bots to make initial patient mental health assessments, and later on potentially even mental health treatment. In other areas of healthcare they are experimenting with using AI to spot cancers on MRI scans, with a better hit rate than trained doctors. Even on the very low level, plenty of businesses are already using ChatGPT to generate written content where previously they would have paid freelance writers.
What no one appears to be thinking about, is what all these workers who are replaced by AI, are going to do for work afterwards. In the UK there are around 33 million people in paid employment currently - if even 10% of those are replaced by AI, that's another 3.3 million unemployed - a two and half fold increase in unemployment over what could be a very short time, a two and half fold increase impact on unemployed benefits the state will have to find funding for, plus the reduction of economic spending capacity of the working population - and every developed economy could see the same globally. The economic consequences of that reduction alone would be crippling.
Personally I think even 10% could be a conservative assessment if the proliferation of AI continues unchecked by government legislation. If you sit and think for a moment about every service orientated role in the workplace (public and private sector), there aren't many that couldn't conceivably be replaced by appropriately trained AI and automation.
AI could well spell the end of humanity, but not with nukes on Judgement Day, but by simple and rapid obsolescence.
