AI or Eh Why?

­Like most people interested in technology, I have been keeping a close eye on AI, especially generative AI, which is intended to be used for content creation. The number of articles written on (and by) AI recently has been astronomical, but how much of it is hype and how much is substance? As a content generator myself, I’ve had a good look at the capabilities of most of the available generative AI apps. Generally, I’ve found their output very flat and repetitive. I was hoping that I could use AI to form the basis of articles to save time, but I can honestly say by the time I’ve edited the copy, checked the ‘facts’ in the article and then finished it, it has taken longer than what it would have if I’d written it from scratch.

It seems that I’m not the only one that is sceptical of its usefulness. Recently, I’ve come across a couple of articles that also cast doubt on its ability to change the world, as has been claimed from many different quarters. The first article is especially relevant here, as it focuses on AI’s power consumption. In the article, Alex de Vries, the founder of Digiconomist, estimated that the power used to create the Google AI snapshot provided in a search result was around 3W, around ten times the power used in a normal Google search. Google announced in May that it would be adding AI snapshots to every search performed on the platform, which de Vries then calculated could consume as much energy as the entire country of Ireland. Considering the questionable value of the snapshot results, along with the fact that it is a distraction from other, more useful websites, and that seems a lot of energy to burn for no discernable advantage.

The second article I read, or to be more accurate, piece of research, came from Goldman Sachs and it looked into the money is being currently spent in AI installations and the benefits the companies that were spending that money were receiving, or at least expected to receive in the future. Their findings were mixed, but is anything less than a glowing endorsement expected considering $1 trillion is forecast to be spent on AI capex in the coming years?  

In the research, perhaps the most damning was Daron Acemoglu, who is an Institute Professor at MIT. He said, “Given the focus and architecture of generative AI technology today, truly transformative changes won’t happen quickly and few—if any—will likely occur within the next 10 years”. He was echoed by the Head of Global Equity Research at Goldman Sachs, Jim Covello, who quoted, “AI technology is exceptionally expensive, and to justify those costs, the technology must be able to solve complex problems, which it isn’t designed to do”.

Other analysts quoted in the research highlighted that spend was not particularly high compared to revenues, but struggled to point out where any returns would come from.