Those who have read my posts in the past know my opinion of AI. It just doesn't exist as an entity, per say. Rather, it's a technique to program computers, much like the object-oriented approach, or the procedural approach.
The 'AI approach' does have some advantages in dealing with 'fuzzy' datasets, and while not perfect, does do a decent job of both decoding and generating speech. It can also be a great tool for manipulating images and video.
But relying on any decisions it makes? Hey, if you are planning on doing that, I have some land in Florida I'm trying to sell.
First of all, any program is driven entirely by the data that it is fed. We all know the accuracy of data from virtually any source, whether it be from news reports, social media, books, people's opinions, etc. You have to take everything with a heavy grain of salt, and look not only at the data, but who is writing it up and even why.
And then there's the decision-making. Deciding what might be close to accuracy, and then coalescing that data into some form of conclusion. Why would anyone rely on anything other than their own intelligence, experience and knowledge? Why rely on what some unknown programmer coded into the system, and then supposedly (and I do mean that very strongly), is manipulated by some other unknown technique (again coded by a programmer), that is mistakenly called 'machine learning'.
Would you stake your life on such decisions? How about your view on certain politicians? I wouldn't even risk wasting money on a restaurant meal recommended by so-called AI software.
It's intellectual laziness. Now that the vast majority of the population doesn't know how to spell (relying on spell checkers), doesn't understand basic grammar (again, relying on checkers), they now want this AI software to run the rest of their life.
There already is a stupidity crisis, and it's not just in America. AI (I really do hate that term, we don't even know what human intelligence and consciousness is, never mind artificial) is just going to exacerbate it. Our brains will atrophy. There will be 10,000 video channels out there, all running reruns of Survivor and The Bachelor.
The only positive I can come up with out of all of this is that maybe, just maybe, some other animal will evolve to the point where they can surpass humanity (assuming, of course, that we don't kill everything off in a nuclear holocaust) and do it right this time.
Is AI simply a form of automation?
Yes. But it's not sold that way. Nor is it used that way by gullible individuals. It's like believing everything you read. Or watch.
Earlier in this thread Artisan Radio linked to an article titled 'Stupidity Crisis" and if you didn't click and read that article I'd like you to go back and read it. It's very seriously true and we're more cowed by the stupidity around us than we realize. To live peacefully with others in present day human society it is necessary to be nice about the stupidity we encounter. It has become taboo to attempt being honest, truthful, frankly outspoken, and offer corrective factual 'setting straight' by way of informing and educating. "Let sleeping prejudice lie" might be our double entendre.
Carl had an interesting question a few posts back, comparing AI to automation, and I probably replied too hastily.
I've been pondering it since, and have come up with what I believe are the differences between traditional automation, and AI.
Automation has typically been used to make physical tasks much easier. Whether it be sorting paper forms, or drilling screws in automobiles, it partially or completely removes the repetitive and tedious task loads off humans. Sometimes, but not always, it does things faster. Sometimes, but not always, more error free.
AI, however, is being marketed as automation to thinking. With few or no disclaimers as to the dangers lurking behind it.
Now, I don't know about others, but I'm not prepared to let AI or anyone else, for that matter, do my thinking. I'm not going to let them make decisions on what I read, how I determine the validity of the facts in what I read, and then make conclusions. That's what makes me, me.
And never mind that. There are so many unknowns behind the facade of AI. Unknown data. Unknown data validation techniques. Unknown reasoning algorithms. Unknown programmers. Unknown privacy. When you use AI, it's like going up to an unknown person on the street and asking them what you should do with your life.
It's already been demonstrated that these AI programs have bugs and make mistakes. Lots of bugs and lots of mistakes. Just like other computer software. Sometimes obvious, sometimes not. They make things up, i.e., they can manipulate the facts. Legal briefs have been filed with these mistakes, scientific papers have been published with these mistakes. People have been serenaded with gobbly gook.
The real danger is not that these mistakes are being made. Even traditional automation machines will have faults. Failsafes and testing are built in. The history of software and programming is littered with failed products that just didn't work, and programmers that wrote excessively buggy software.
No, the real danger is that AI is being marketed, as I stated above, as thinking automation. If you read the article I referenced in a previous post, and also referred to by Carl, you'll find the formal definition of 'stupid' as not using whatever your intellectual capabilities are to the fullest. With AI, the 'stupid' people are celebrating.
Based on my reluctance to accept AI as a mature and trustworthy tool, I can see a place for it as a toy. For example I would like to have an on air staff of DJ announcers, perhaps the morning shift could be hosted by one human and one AI announcer. the back and forth banter between the two could take interesting twists by being unexpected and unpredictable. And to mix things up the AI guy could be appointed program director to see what kind of goofy results would arise.
Looking far afield I wonder if an AI president of the United States might just be a tad better than what we've been getting.
And here's where I part company with a lot of 'believers, but I suspect that an AI led religion would resemble completely any of the human religions we already have.
May AI prove me wrong.