2024-09-04
AI is a marvellous tool with many uses, but especially perhaps at this time for the high-level scammers in our world - how so?
Almost nobody (yet) understands how it works!
So we are all fodder for the scam-artists who want us to believe that AI will "rule the world", "take over all our jobs", "always win on the battlefield" (except maybe against another AI!), be all-knowing, etc etc.
So "is AI really omnipotent and can it replace human beings" is a potentially existential question for us.
That AI has its uses is incontrovertible - that it could completely replicate the functionality of a human being is unproven as yet. Some think it might - but as an IT professional of long standing I remain in the unconvinced camp.
Why?
Because traditional digital computers are designed and built by humans and operate entirely as designed by us. If we can't even repair ourselves when we fall sick, how will a computer designed by us ever achieve such a feat?
Do we even know how the human body works? We have been looking for a "cure for cancer" all my lifetime and seem no nearer to that goal today than when I was born - some think we have moved further away, with health issues never known in my youth now proliferating unchecked, notably autism.
There again, some believe that "Big Pharma" long ago mislaid its moral compass and now operates to a commercial paradigm of keeping people sick and marketing them palliatives for "side effects"; whatever the truth of this, it seems to me that it doesn't bode well for any AI designed by humanity to outperform humanity - it might however extinguish humanity ...
(I'm leaving "quantum computing" out of this discussion because I understand next to nothing about it)
Anyway, the point I want to make is that there is not just one kind of AI, there are many, and more I suspect still to be invented.
For instance, technically the Statute Book can be regarded as an AI - it's a set of rules and regulations that Parliament has decreed must be followed, and is considered by some to be notoriously unresponsive to the reality that exists outside of the official brain-cells. Whilst writing an automated AI to match the specific factors of individual circumstance to determine the appropriate action that the Statute Book might demand could be useful in terms of accuracy of analysis and time to respond, it would still have to work its way through the case-book of legal precedent to see whether the courts might agree with its conclusions ... not theoretically impossible, but we begin to comprehend the complexity. And then what about the cases of ambiguity still to be brought before the courts?
Or would we prefer to be tried and released/convicted by the "Court AI", in which case we would be living within the Automated State - "Computer Says NO" on steroids, so to speak.
So for something completely different, take for instance an AI with the ability to play chess. This kind of AI has been faked(!) developed and strategised for years now, and it looks completely different to the latest not-general-purpose-enough AIs ...
So will the future of AI be an amalgam of many and various AI implementations that will somehow be melded into a single overarching general purpose AI?
I'll wait ...
Meanwhile there seems little doubt that AI in its various and probably ever-multiplying forms will be here to stay, and with these we must be prepared to live.
And that will require ensuring that we turn them to our advantage, and turn them away from our disadvantage.
That assuredly will pose difficulty enough for the incoming years.
