With the leaps and bounds in Artificial Intelligence (AI) progress over the last few years it seemed like AI’s future success was guaranteed and our lives would be changed for ever, hopefully for the better, but perhaps that’s not as realistic as we first thought. Or likely to take a bit longer than we thought it would?
It seems that more and more articles are being written about some of the unintended consequences of AI adoption. Two that I came across recently are particularly worrying – here’s my take on them …
Energy cost can outweigh financial gains
It seems that the amount of energy required to produce useful AI is dependent on access to vast amounts of ‘free’ energy, yet we all know that no matter what type of energy is used, it all comes with a cost, regardless of how it is generated.
Data scientist, Alex de Vries estimated in a paper written in 2023 that “… AI servers worldwide could use about as much power as a midsize economy, such as the Philippines or Sweden by 2027 …” (that’s a mere three years away!).
Despite countries trying to meet the current and future energy needs of their own countries by installing renewable sources (China installed more solar panels last year than the US ever has, and India plans to triple renewables in the next six years), this growth hasn’t kept pace with demand, leaving coal, the worst emissions-offender, to close the gap. No matter how much renewable source is added, it can’t keep up with demand which means dirty, emission-heavy coal will continue to dominate energy creation for many years to come in developing countries.
The boom in industrialisation and the demand for basics like lights and air conditioning, along with the growth in uptake of electric vehicles, heat pumps and other devices used to reduce fossil fuel use, seems to be creating a vicious circle of infinity. And the rapidly increasing use of artificial intelligence is only adding to the shortfall. With a surprisingly real cost in dollar-terms of the energy needed to power AI, it clearly won’t be coming free or saving the planet’s resources, as many people might hope. AI technology’s real value is in the time it saves humans to do things.
With the advent of AI, which is yet another whizz-bang techno-tool created as a way to eliminate human effort, comes the negatives – there’s always a downside, isn’t there? The sheer amount of energy consumed probably wasn’t factored in at the beginning, and as happened with crypto currency mining, the effort seems like it’s going to be greater than the returns, at least in terms of energy consumption.
Truth and authentication matter
Another unintended consequence is how it’s used. Apparently, people don’t want to see AI created images and text in marketing campaigns, and yes, they can spot the difference. People want the authenticated version provided by a human being. They want the truth.
Savvy marketers are using AI to provide new ideas, new uses for products, providing two outcomes or versions of a concept, or even simple things like tweaking colours or formulae. In other words, it’s being used to assist behind the scenes in a wholesale way, not as the final or end result, though it can be used successfully to create first or second drafts as well, as long as it’s fed enough background information to get the message right.
Once a campaign has been created, AI becomes invaluable by building it into other applications, such as on social media platforms like Instagram or LinkedIn, or as videos or podcasts.
That’s great as long as it doesn’t sacrifice ethics and integrity along the way. It’s been used to cheat in exams and assignments, manipulate decision making, and create impressions it might not be able to deliver on. Currently its being used by scammers to cheat people, as it produces much better-composed begging letters than the offenders would ever be capable of creating on their own (good syntax, no spelling or punctuation mistakes, and no poor grammar) which makes it far more difficult for intended victims to quickly identify a scam.
AI’s contribution to the economy through increased GDP
Amazon founder, Jeff Bezos, once said that AI is the great equaliser, however while it might equalise your skill set, we’re still looking at ways it can equalise your ability to take advantage of prosperity as well. GDP (gross domestic product) has increased all over the world, however there are only a few companies contributing to the GDP increases in the USA. Why is that? Is it just a timing issue or how it’s being used?
While it might be early days for AI to reach its full potential, there’s no denying it’s a game changer in so many ways. Whether the amount of energy expended on AI to produce a desired result is a good investment might be debateable. Is AI just providing another means for evil to overcome good? How come it’s so slow to add value to an economy – who is really benefitting financially from AI?
Regardless of the answers to the above questions, I’m pretty sure AI is the way of the future and will be life changing in so many spheres of our lives, but we might just have to wait and see how it unfolds.
Remember that the motor car was invented to reduce the pollution created by horses in the 1800s and look where that has ended up. Unfortunately, unintended consequences exist with every new labour-saving invention created, so we need to be careful what we wish for.