We should beg Google to charge us for using AI

It’s worth paying search giants for artificial intelligence to protect our data and avoid being bombarded with advertisements

It remains to be seen whether Google introduces charges for its artificial intelligence (AI) products, and, if it does, whether the other emerging giants of the technology follow its lead. Perhaps every question will cost a few cents. We might have to pay extra for longer, more detailed answers. Or there could be a monthly subscription option, covering everything you might possibly want to know.

One point should surely be absolutely clear. As consumers we should be begging Google to make us pay. Sure, free stuff is great to start with. But it also means everything is driven by advertising until it becomes unusable.

Paid-for AI will be far better, and much more useful in the long run – and it is customers that will come out ahead.

It could turn out to be the biggest change of strategy in the company’s 25-year history. According to reports this week, Alphabet, as the owner of the Google search empire is now called, is weighing up whether to put it AI-powered search features behind a paywall, forcing consumers to pay to use them; or whether to keep them completely free, as its basic search function, email server, maps, and all the other services it provides are at the moment.

Other AI services are grappling with a similar dilemma. ChatGPT, currently the runaway market leader, has a free version, and a slightly superior paid-for model, and most of the other major AI tools are opting for something similar.

This is a tough decision for Google. After all, giving stuff away for nothing has worked out brilliantly for the company. It has built a business worth a staggering $1.8trillion (£1.4trillion). Its advertising revenues are $175bn a year, mostly from its search engine and other related products. It has been a winning formula, and it will be very difficult for it to change now.

Against that, however, AI poses two major challenges. We may all gradually stop searching for information and products on the internet and simply ask a smart chatbot to do it for us. If that happens, Google’s search ads will steadily become less valuable.

Even worse, AI uses so much server power, and such sophisticated chips, that it is far more expensive to run than traditional web pages. Caught between those two forces, Google may easily find its profits start to get squeezed.

We all like getting something for nothing. The explosion of web services over the past 20 years means that we all get a whole range of incredibly sophisticated products effectively for nothing.

We can search for anything we want, send emails around the world, find anything on maps, chat to friends and family on social media, and store our photos for ever and, apart from a broadband connection, it doesn’t cost us anything at all. It is, in some ways, a great deal.

The catch was that it all had to be funded by advertising. Over time, search became more and more useless as the answers to any question were dominated by sponsored results. Mail systems were cluttered up by messages from companies, and maps were dominated by nudges towards one shop or another.

Meanwhile everything we said on social media was bundled up and sold as a commodity to be traded by marketeers.

As a former Google web designer memorably put it: “If you are not paying for the product then you are the product.”

AI could very quickly go the same way. The chatbots will start casually throwing in product recommendations to every conversation. They will hint at different shops or restaurants we might want to visit. They will slowly work every question around to a list of things we might want to buy.

Worst of all, they might start guilt-tripping us into donating every time we log on, like Wikipedia, or indeed The Guardian. And then of course they will keep a record of every conversation and question, analyse your answers, and sell the data to squeeze some more revenue out of you.

That is what the tech giants such as Google and Meta, the owner of Facebook, Instagram and WhatsApp, did very successfully with the first version of the internet, and it will be very tempting for them to repeat the trick all over again with AI.

If they get it right they could rake in even more billions in revenues.

It would be far better to pay a small monthly subscription and get a better product. If we don’t, AI will go the same way as the rest of the internet, and we will all end up as the “product”.

Smart chatbots are expensive to run. It is estimated that ChatGPT already costs $700,000 a day just to operate, and that is before you add in all the development costs.

As AI systems start paying for the content, as they should, and as they get quicker and smarter, as they will, they are going to cost more.

It was a mistake to make everything free the first time around, and there is no point in simply repeating that error all over again.

It would be far better if everyone was charged a few pounds every month and the product got steadily better – and was not cluttered up with advertising.

True, we might be reluctant to pay any charges to start with. We have become very used to anything digital being free.

But given the potential of the technology it is loose change. We should be begging Google, and the rest of the AI giants, to charge us. We will be far better off in the long run.

‘Paid-for AI could turn out to be the biggest change of strategy in its 25-year history’