Opinion – Ronaldo Lemos: Will creative jobs be replaced by artificial intelligence?

It was long believed that artificial intelligence would do away with jobs of a repetitive nature. In a famous study published in 2013, Carl Frey and Michael Osborne argued that the more repetitive the nature of work, the more susceptible it would be to automation. They listed workers in maintenance, transportation and machine operation services as highly likely to lose their jobs in the coming years.

Creative workers, on the other hand, a segment in which complex intellectual operations are involved, such as artists, musicians, photographers, writers, designers and others, would have a lower probability of automation.

We are now seeing the opposite of what Osborne and Frey predicted. It is the creative workers (and the cultural industry) who are at risk of cataclysmic change. The reason is the rise of publicly accessible artificial intelligence models that are capable of “creating” virtually any type of content with a few text commands, either for free or for significantly less.

These models include ChatGPT, Dall-E, Stable Diffusion, Hugging Faces, AlphaCode, and so on. The capacity of these systems is both impressive and frightening. In ChatGPT, it is possible to generate any type of text, as well as to search for information previously impossible to obtain in an automated way.

For example, I asked ChatGPT which culinary dishes are mentioned in Jorge Amado’s books. At the time, he knew how to identify these categories and produce a coherent list: feijoada, moqueca, vatapá, acarajé, shrimp bobó. I could have asked the system to write any type of text or essay, and the answer would have been coherent in most cases.

ChatGPT and similar models can replace an overwhelming amount of human creative work. Be it in the generation of journalistic texts, be it in the production of content for social networks and newsletters, be it in the service in call centers, and so on. And, of course, the automation of malicious activities such as creating fake news, inflammatory content or cheating in the educational world. It’s going to get harder and harder for a teacher to detect plagiarism from students using ChatGPT to do their assignments in any subject (math, biology, geography, etc.).

The problem doesn’t stop there. Artificial intelligence models are capable of generating any content. Excellent quality photos, drawings, images, music, movies and animations. For example, one of the models is already capable of generating playlists on Spotify. You type in the kind of mood you want (“New Year’s Eve Party with only upbeat Brazilian music”), and the playlist pops up instantly.

On the one hand, this democratizes skills that were once specialized. On the other hand, it has the potential to destroy the value of creativity. Why would anyone hire a photographer or buy images from an image bank if I can generate everything for free using artificial intelligence, with sometimes better quality than “real” content?

This also raises questions, including regulatory ones. Brazil, for example, is discussing a bill to regulate artificial intelligence. He was born obsolete.

It was made based on other concepts and concerns. It is practically incapable of dealing with the challenges of these new models that are among us. Theme for another article.

READER

Already it wasemoticons
Alreadyemojis and stickers
It’s comingAI-made avatars able to speak and express emotions