Opinion – Bernardo Guimarães: Artificial intelligence, economic progress and inequality


The development of artificial intelligence has disputed space with the World Cup in social networks and conversation circles.

Not long ago, it was images generated by artificial intelligence that impressed with their style and creativity.

In recent days, the subject has been the ChatGPT tool. Recently released, it is able to create stories from user suggestions and provide often reasonable answers to various questions.

Since it was launched, several problems and limitations of this tool have been pointed out by humans. There’s still a lot of room for improvement.

Still, several texts and responses are impressive and suggest that these new models based on artificial intelligence could be a relevant part of the work done today by humans in the near future.

This raises a number of questions (ethical, philosophical, psychological…). I will focus on economic issues.

Technological innovations have two types of effect.

The first is that society as a whole can produce more or work less (or a combination of both).

Think of a family that likes to make illustrations for their own consumption. A tool that creates illustrations replaces human labor. The family can use the tool to make more illustrations or they can use the spare time to produce other things or for leisure.

This effect continues to exist in a world where people work producing illustrations for others. With new technology, we can have more illustrations or fewer people producing them (or both). We do more with less. Looking at the whole, this is good.

The problem is that, in this case, fewer people producing illustrations means that those who worked with it lose their jobs or have their income reduced. A skill developed with years of study and work loses value.

This is the second effect of technological innovations: they affect the market value of different occupations and therefore the distribution of income.

In recent decades, technological advances have replaced routine tasks, making lower-paying jobs less valuable and more specialized jobs more valuable. This was a key factor in the rise in income inequality in developed countries.

Ideally, public policies would encourage innovations to take advantage of their benefits and help those who lost with advances to minimize the negative effects.

This, however, is difficult. There are ways to reduce inequality or compensate for those who lose out from technical progress, but these are costly and imperfect. Therefore, there are those who defend policies that effectively reduce the pace of technological progress that replaces low-paid jobs. This question is important and complicated.

Another implication of this discussion is that public policies should encourage the formation of people capable of producing in a world with these new technologies.

The advancement of artificial intelligence brings these issues to the forefront, with some nuances. One is that, with the pace of progress in this area, it is very difficult to make predictions for 20 years from now.

Which occupations will be left to humans, which will be replaced by machines? What skills and knowledge will be required? Which ones will become less useful?

Trying to find clues, I asked ChatGPT to write poems and talk about their limitations. With some prodding, she replied something like this: “I’m a poet and I didn’t learn to love, I can’t feel what you feel when you read a verse I wrote”.

You May Also Like

Recommended for you

Immediate Peak