Aardvark Daily

Please visit the sponsor!

Should we really have commercialised AI?

It's fitting that one of my final columns of the year is (once again) on the topic of artificial intelligence.

Fitting because AI, no matter how you look at it, is going to change the world. We can argue at length over whether the changes created by AI are for good or bad but it is undeniable that there will be changes; significant changes.

There are several scenarios that play out and I'm afraid that from where I sit, very few of them are entirely positive. In fact, some are entirely cataclysmic in scope and magnitude.

First, let's look at the positive.

When applied to many of the challenges and problems associated with science and technology, AI has already shown an incredible ability to outperform man in some areas. It can design and model new medicines, come up with solutions to complex problems that have evaded our best efforts and provide amazing insight into understanding some of the greatest scientific mysteries.

These applications of AI could significantly boost our ability to advance our understanding and use of technology in the modern world, for the benefit of all.

Unfortunately, that's about where the good news ends.

The problem, as I see it, is that AI is being commercialised to such an extent that it will change the world's economy and the structure of our society in far-reaching ways that do not seem to have an overall benefit.

AI is already taking jobs, as has every other technological revolution before it.

The industrial age saw factory labourers replaced by automated production lines, the information age saw office workers replaced by computers and the AI age will also see huge swathes of workers and creative-types displaced by large-language models and generative AI systems.

The problem is that unlike the previous tech revolutions, the AI age won't be creating new jobs to replace those lost. The huge numbers of people who find themselves redundant may not be able to find employment and that's a problem -- for all of us.

What the AI age will do is see a massive transfer of wealth from those who were previously "workers" to those who stand to benefit from the power of that AI. Those beneficiaries will be company owners, stock-holders and others that see their incomes climb through the reduced costs associated with replacing wetware with AI systems.

Rising levels of unemployment place a burden on the welfare system and that is funded by tax -- yet we all know that the truly wealthy have quite effective methods of sidestepping much of the tax net. We could see governments forced to run huge and unsustainable deficits and resort to significant borrowing, just to keep people from starving.

Those whose businesses have embraced AI and shed boatloads of workers ought not be too smug though. Although they may see significant upticks in profit in the short term, the reality is that if huge swathes of the population are unemployed and facing poverty, there may be little to no market for the goods and services that are being boosted by AI.

The end-result could be a financial collapse unprecedented in the history of the Western world.

AI may, in effect, result in the destablisation and destruction of the social and economic models we've built over centuries.

There is also another negative scenario and that's the AI bubble.

Right now, many billions of dollars have been and continue to be invested in the development and deployment of AI systems. Many companies and many nations are throwing everything they have into becoming the "leader" in AI tech. To not have a horse in this race is seen to be putting your country and its people at a huge disadvantage.

So what happens if AI fails to deliver on the promises made for it. What if all that investment does not produce the hoped-for returns to those investors?

Anyone remember the dot-com bubble? When investors finally figured out that none of these internet companies were actually making any profit they began to dump their shares and the value of the market crashed almost overnight. The Yahoo! stock that made you a paper millionaire one day was nothing but a liability the next.

Well the AI bubble is much bigger than the dot-com bubble so if it bursts the fallout will also be much bigger and it could plunge the world into a significant recession that would impact everyone.

Finally, one oft-overlooked aspect of all this AI investment is the environment.

AI datacentres are using horrendous amounts of energy, some of it from non-renewable sources so that it contributes to climate change. Even worse however, is the fact that the hardware being used (processors, NPUs, GPUs, etc) to run all these LLMs and generative AI systems, is advancing at such a pace that most companies have to depreciate their investment over just three years. Within that period they need to throw out all the existing gear and replace it with new stuff, so as not to bear the unsustainable burden of slower, less energy-efficient, less powerful architectures.

This represents billions of dollars of hardware that will have to be scrapped or thrown on the recycle heap. Do we have the capacity to cope with such a mountain of industrial ewaste?

Let's not also forget that even the cooling for these datacentres can have a decidedly negative effect, consuming valuable water resource and creating microclimates that never existed before.

Viewed from these perspectives I tend to think that although the use of AI for scientific and technological reasearch and development is great -- using it to replace human workers may end up causing far more trouble and pain than it's worth.

Sadly, the head-long rush to be "the winner" in the race to create the best AI systems means that caution has been thrown to the wind and little consideration is being given to the potential cataclysmic side-effects that may result.

Or I could be completely wrong -- however, it is food for thought.

Carpe Diem folks!

Please visit the sponsor!

Here is a PERMANENT link to this column

Rank This Aardvark Page

Change Font