April 15, 2024
A.I

Technology will replace jobs. That’s exactly what we need

This is Armchair Economics with Hamish McRae, an exclusive newsletter for i subscribers. If you’d like to receive this straight to your inbox, every week, you can sign up here.

Artificial intelligence is the hot topic at Davos this week. The companies that pay for the World Economic Forum’s annual party in the Swiss Alps are pushing hard for it to happen. There is a huge poster of Tata Consultancy Services proclaiming that “The future is AI”; one of the large buildings has been temporarily renamed AI House; and several high-tech giants, including Intel, have placed AI slogans on their bases for the event.

Because? Well, the basic answer is that this is where the global business community believes it can make the most money over the next decade, and no one wants to be left behind. In recent days, Microsoft’s partnership with OpenAI has helped it overtake Apple to become the most valuable company in the world. Its founder, Bill Gates, who is now not involved in the company, just conducted a fascinating interview with CNN’s Fareed Zakaria, in which he predicted that AI will make everyone’s lives easier.

For example, you could help doctors do paperwork, which is “part of the job they don’t like, we can make it very efficient.” He also said the improvements with OpenAI’s ChatGPT-4 were “dramatic” because it can “essentially read and write.” Then it was “almost like having an administrative worker to tutor, give health advice, help write code, help with tech support calls.”

The traps

Of course, there are scary stories that come out galore. The latest is that AI can not only clone people’s voices and create fake videos, but it can also imitate their writing. The traditional device used to verify identity, a signature, seems in danger; The newest one, which involves using your voice as a password, probably won’t last much longer. As for fake videos, the only way forward I see is to increase the penalties for the people who create them. But remember, AI will help identify criminal activities, as well as provide new opportunities for offenders.

Whenever a radical new technology appears, it takes a while to understand all of its economic or social consequences. But I think now that we’ve been experimenting with ChatGPT for over a year (it launched on November 30, 2022), we’re starting to get a little clarity on its likely economic impact. Here are three general areas where that is starting to happen.

Administrative assistance

One will be the elimination of repetitive administrative tasks. Bill Gates points out how the time doctors need to prepare medical records can be reduced. There are similar huge parts of administration that we all find tedious. This month there are 5.4 million Britons due to file their tax returns by January 31. If AI could act as a tutor, sitting next to people while they did it, think of the time saved and the distress alleviated.

Viewed from the government’s point of view, there would be fewer errors, fewer penalties, and revenue would arrive more quickly. It would also allow them to offer better service. Apply that to all public sector activities and you can see the potential. Planning applications could be processed within days; hospital waiting lists would be reduced; court delays, etc. could be eliminated. Nobody likes filling out forms. Of course, things have to be checked, but AI could do the legwork.

Educational benefits

The second is education. I’m not talking about teaching as such, nor the more contentious question of whether students should be allowed to use AI to contribute to their work. We cannot yet pass judgment on these issues. But what has become clear is that it can help reduce the workload of teachers in schools, and perhaps also in universities, when grading and evaluating student performance. You have to check everything again. But if the time needed to mark essays could, say, be cut in half, it would free up time for more constructive interaction. Schools could do a better job.

There is a lot of research being done, summarized in this article from Harvard University, and the conclusion is that this is a tool that will revolutionize education for the better.

Fake news

A third is the information industry. For those of us who work on it, it could seem like a threat. Misused, it could be. There is currently a legal dispute over copyright, with The New York Times suing OpenAI and Microsoft for inappropriate use of their material. However, Sam Altman, director of OpenAI, told delegates in Davos that he does not need large amounts of training data from publishers like the NYT. What they need are smaller amounts of high-quality data, and they are striking deals with other publishers like the Associated Press and Axel Springer to get it.

So I think we are in the early stages of establishing a relationship between companies that use AI for information purposes and the publishers of the content. If done right, AI could help crack down on fake news, and all decent journalists should be grateful for that.

It’s complicated and exciting. But I predict that within a decade we will have come to consider what is happening now as important as Steve Jobs’ launch of the iPhone in 2007. If you are reading this on your mobile phone, you will know what I mean.

I need to know

However, there is a conundrum with the introduction of new technologies, and the iPhone example highlights this.

It is that, in fact, our lives have been completely transformed by mobile telephony, but there has been no significant increase in living standards in most developed countries in the decade after 2007. (It is better to consider what happened before the pandemic as the numbers since then are all over the place.) Take the United Kingdom as an example. There is a report from the Resolution Foundation that points out that, although average income increased a little, average income was practically the same in 2017 as in 2007.

Living standards

However, if you look at everyday life with the large number of applications that clutter our phones, that cannot be correct. Personally, I would find it very difficult to keep friends and family in touch without WhatsApp, and I am quite averse to social media. So there must have been an improvement in our living standards that does not appear in the GDP per capita statistics. Various efforts have been made to estimate this, for example by asking people how much they would pay to retain access if it were taken away. But I’m not sure you’ll really get sensible answers because the counterfactual is always difficult to evaluate and in any case we are all different in our habits and choices.

However, and this is my key point, I think in the case of generative AI it will be different. Its impact will be reflected in living standards figures, partly in falling costs and partly in higher quality of services. If waiting times to see a doctor can be reduced, this represents a real improvement in both the standard and quality of life. If it increases longevity, even better. How long until we start watching this show? My guess is the next two years. Let’s hope we’re right.

This is Armchair Economics with Hamish McRae, an exclusive newsletter for i subscribers. If you’d like to receive this straight to your inbox, every week, you can sign up here.

Leave a Reply

Your email address will not be published. Required fields are marked *