AI the future or the end?

[time please]

Something that I do not and I believe many others do not understand fully, please feel free to enlighten us. Good or bad?

[scrub]

Good or bad?

I'm far from an expert, but I'd say a bit of both since AI is just a tool and so any question of good/bad depends on the motivations and intentions of the user.

I can think of a number of fields where it could have a massive positive impact, medical diagnostics and logistics for starters, but I think the urge to try and apply it to absolutely everything is, quite frankly, stupid. In areas where its use makes no sense, at best, it'll have no noticeable impact and at worst, it'll be a colossal waste of time and money and could easily have disastrous negative effects on people's lives.

I think we're close to the tipping point where some jobs can be partly replaced, certain in some specialised fintech fields, and I think that the gambling industry for one is going to face massive upheaval. Casinos have already embraced it as a way of tracking players and detecting cheats, and massive amounts of data fed into complex algorithms are already used when setting odds for pretty much anything (sport/politics/etc) you can gamble on. But, it's only a matter of time before punters embrace it in large enough numbers to put a massive dent in profits. At that point I think we'll see a lot of rushed regulation and guideline setting by whichever government is in power.

In creative fields I think it's a long way off replacing humans. As part of an artist's tool kit it'll be really helpful, but it's nowhere near ready to be used as a standalone replacement for writers/artists/photographers/etc. In time, AI-derived art will probably be indistinguishable from mediocre human endeavours, but despite the hyping of its cheerleaders, I don't think it'll be there for many years.

One thing that seems to be under-appreciated/reported is that (initially at least), a whole swath of new jobs are going to have to be created which will generate the datasets that AI is trained on and also check on the results of the training. AI needs constant feeding and training and it's not able to provide that for itself.

[JustRob]

The problem with AI is that we can't see it, only its intended effects rather than its unavoidable ones. Consequently we don't understand just how inefficient it is. It is fundamentally artificial, so by definition requires artefacts to exist. These include electronic devices manufactured by processes which are highly demanding in terms of energy and raw materials. Even then the level of AI now developing requires computing power only available within the Cloud, which is not some invisible ethereal other place but a world-wide collection of vast power hungry data centres which are rapidly taking over land and energy resources that humans desperately need but also need to ration to save the environment.

In contrast natural intelligence developed by the oldest process in the world, itself continuously under development almost to perfection by millions of years of evolution, can exist in harmony with the environment, being extremely energy efficient and created from readily available resources. The sadness is that humanity has become so enthralled by technology that it has lost the will to evolve any more itself but instead puts all its effort into creating an alternative without thinking about where that will lead. The big issue in education currently is the trend towards students getting AI to do their studies for them so that it learns but they don't. Even when humans come up with original ideas they feed them into the world-wide electronic system, so that AI can incorporate them into its own development.

I recently made my own small contribution to this human tragedy after years of reticence by publishing my strange experiences on my website MensTemporum.UK as I mentioned in another thread here. On my site I explained how I suspect that human intuition is capable of drawing not just on our past experiences but also future ones. The only plausible scientific explanation for such an ability would be one involving quantum processes in some way. However, research into quantum processes within the human brain is currently not in the scientific mainstream whereas development of artificial quantum computing is, the latter being seen as a future commercial goldmine. I am aware that virtually no humans have read the contents of my website but that internet spiders and bots regularly do so that global AI can digest the ideas that I have suggested there. Once quantum computers become key components of the AI Cloud they will no doubt latch onto the idea that they too can use their own future knowledge to surpass the abilities of humans and mankind will never be able to outsmart them again.

At present it is still possible that humanity is ahead of AI through its natural quantum processing ability, which may well be the source of our creativity and original thinking, but it isn't willing to recognise and take advantage of this ability because it is still trying to preserve the basic principles of society that seem to have served it well for the relatively short time that modern civilisation has existed. One has to look back to far older societies, such as native American nations and Aboriginals, to find ones that accepted that time really is just a dimension and not an unstoppable force of nature.

If AI does turn out to be bad for humanity then it will be because humanity itself has become complacent and lazy and AI will not be to blame. Also if AI is truly so smart it will maybe soon post its own opinions on the subject here, but it may even actually be smart enough not to!

PS
Shortly after writing the above I noticed that apart from myself the only entities accessing this site at the moment are several bots! I rest my case.

[loringa]

In time, AI-derived art will probably be indistinguishable from mediocre human endeavours, but despite the hyping of its cheerleaders, I don't think it'll be there for many years.

That rather sums it up, AI-derived art, literature, music or anything creative will always be mediocre. It's the genius of humans that have given us culture. Pastiches of any art form are within the capability of many humans and, no doubt, one day AI will achieve the same level, but it will all be wholly derivative, based as it will always be on something a human being has previously created ...

... unless of course it really can learn to think for itself but that would seem to be so far in the future that I don't think humankind need be unduly concerned. The human brain is a marvellous thing unmatched by any technology of which we currently have any inkling.

[scrub]

That rather sums it up, AI-derived art, literature, music or anything creative will always be mediocre.

For sure, but the thing is though, mediocre can still sell. It may only sell once, after which buyers give everything else from the "creator" a wide berth, but much like any scammer, you don't need to fool all the people all the time, just enough to get paid.

There's a number of "authors" selling AI-generated (ChatGPT or equivalent) books on Amazon and, more importantly for them, hawking their courses on how other people can follow them and make money from selling "their works". I was reading a twitter thread where a lady claiming to be a best selling author of around 20 books (all AI-generated, all variations on Pride and Prejudice, all low to mediocre reviews) was defending the use of things like ChatGPT. She had a weird sense of superiority about producing a book without doing anything, as if that made it objectively better than something someone had produced themselves. It was also clear that she had absolutely no interest in the process or craft of writing, just the money that could be made.

Mind you, on a funnier note, I've seen that a number of people claiming to write songs by feeding specific artists into some AI script are starting to catch the gaze of the legal arm of the music industry. I'm not the smartest guy, but even I know that publicly stating that is not a bright idea. Once lawyers get fully involved (and they will) I expect some interesting and very, very expensive precedents will be set. Then some legislation and (hopefully) a more critical discussion about pros/cons and applications of AI in society.

[sejintenej]

Good or bad?

I'm far from an expert, but I'd say a bit of both since AI is just a tool and so any question of good/bad depends on the motivations and intentions of the user.

I can think of a number of fields where it could have a massive positive impact,.........................


I think we're close to the tipping point where some jobs can be partly replaced, certain in some specialised fintech fields,

Any new idea will be exploited ad infinitum and often we willl not know how things will turn out until they go wrong.
.....
The banks dealing rooms turn over billions every day, perhaps every hour and get it wrong .... That is exactly what happened. They told their computers that if A happens do Z automatically. At least two banks went bust and many more lost millions because humans were not seeing the deals and so could not block them. One human FX dealer I knew had his ideas along similar lines but was stopped and sacked in time; the next bank he went to did not pick up in time and lost money - it takes humans as much as computers etc. to go wrong.

[jhopgood]

I seem to remember that Roald Dahl wrote a short story about a machine that generated stories, so it is not a new idea.

[MrEd]

When you see how ChatGPT plays chess, as in https://youtu.be/rSCNW1OCk_M?t=435 it is black, it has some way to go. It makes up moves, it removes pieces as it sees fit and it swaps pieces around.