Linguists, biologists and publicists, the new experts setting up as ‘prompts engineers’ for ChatGPT

Before generative artificial intelligence turned his world on its head, Guillermo Prado did a PhD in molecular biology in the oncology department of La Paz, a public hospital in Madrid. He then started working in a company related to the healthcare sector, but he didn’t enjoy it. It was late 2022, and the first versions of ChatGPT and Midjourney, an application capable of creating images from text, were released at the same time. “I started going crazy with it,” Prado says. Now he is an expert in the new skill known as “prompts engineering.” Salaries in the field can reach six figures in the United States.

“I thought that was the future, so every day when I came home from work, I started researching it,” Prado says, embarking on a path of no return that has led him, two years later, to create Externia, his own events company driven by generative artificial intelligence — technology capable of creating text, images, music, or videos from written instructions. In the face of the caution and general lack of knowledge that still dominates public conversation, Prado believes that the revolution led by this technology will be transversal, affecting all jobs and opening the door of innovation to professions that have historically been left out of technological progress.

“Why would a humanities profile work as a prompts engineer?” asks Prado. Prompts are the instructions or requests received by the machine. The prompts engineer is the expert professional, for example in ChatGPT, who knows how to get the most out of the tool by making use of language. “The profiles that are coming in now are usually computer scientists, and that’s fine and it works, they’re going to be needed. But the most important part of a prompts engineer is to be creative, and there I think humanities profiles [such as philosophy, marketing, or linguistics] can have a lot of potential.”

“When it comes down to it, you are communicating with a natural language, and a profile that has studied that will have a huge advantage,” Prado tells EL PAÍS in a telephone conversation. The connection between the humanities and generative AI seems minimal, but more and more voices are pointing in that direction. Marco Argenti, the CIO of Goldman Sachs, wrote about the subject in the Harvard Business Review: “The quality of the output of a Large Language Model [like ChatGPT] is very sensitive to the quality of the prompt […] Because of that, one would have to first and foremost master reasoning, logic, and first-principles thinking to get the most out of AI.”

Artificial intelligence could alter almost 40% of global employment, according to the International Monetary Fund (IMF). In advanced economies, that figure rises to 60%, because the number of jobs performing replaceable tasks is higher than in other economies. However, a recent article in The Economist asserts that those with humanities degrees are well positioned to take advantage of the new opportunities. “Promotion agencies and online courses purporting to teach this skill are flourishing. Graduates with a background in languages or the humanities are popular candidates,” the text states.

Momentum in marketing

Ruth Falquina, CEO of the marketing agency Estado Latente, studied marketing at the Complutense University of Madrid before entering the world of artificial intelligence. She now uses that technology throughout the productive and creative process of her company. “Humanities profiles bring critical thinking and creativity based on a deep knowledge of culture. This background serves to generate better prompts, better instructions,” she explains. Falquina, like many other experts, says that “prompts engineer” is not going to become a job per se, but a skill that many people will have to develop for their jobs.

At her marketing company, employees use tools like ChatGPT on a daily basis. It’s part of the production process they go through when they have a new assignment. Falquina notes “co-creation” between human and machine is essential. “We integrate AI as part of our team,” she says. They have used it for smaller things, from producing the first AI-based cover in Spain (it was for the first issue of Yorokobu magazine), to creating a drink wrapper thanks to machine design. For the Jaume I scientific awards, for example, they have created an AI that helps the user to lose fear of the AI itself. They have also generated a ChatGPT model to “help strategists around the world develop their own brand.”

“It improves work processes and inspires us to be better creatives,” says Falquina. Another project was the main stage designer for the Mobile World Congress. They had two and a half weeks to complete the final product. “In that case, the AI allowed us the speed we needed to deliver everything we were asked for.” Hermagoras Abecia, a digital marketing specialist at Actahotels, has also been tinkering with ChatGPT for some time, so much so that he has already given himself the “prompts engineer” skill on his LinkedIn profile. He says he has been developing his own tools since October last year, and that his AI has become a sort of companion with whom he converses most of the day.

“You have to know how to give instruction”

“It helps you from doing a context analysis for a marketing campaign to a Google positioning strategy. I do it with ChatGPT and we pull it off,” he says. Another fan, Cristina Aranda Gutiérrez, earned a PhD in Hispanic philology before entering the world of AI, and launches a staunch defense of the humanities in this sector. “Anyone can choose to work in this, because it is something as non-technical as giving an instruction to the machine. But linguists and philosophers are going to know how to do that better than others. In addition, you have to know how to give the instruction, because the context and the precision of the command are important,” she says.

Nico Bour, co-founder of Joppy, a Spanish startup that connects companies with tech workers, says they have not yet seen a huge demand for instructional engineers. “We’re not seeing massive offers for that position. There are some, but no one can tell you yet that they are senior specialists in the field. In many cases what’s happening is that a lot of developers are practicing in their spare time.” Bour believes that the ideal profile for this type of work are data analyst engineers, because they “understand better how an AI works from the inside.” As for the rest of the engineers, “if they don’t learn how to incorporate AI into their work, they are going to be very inefficient a few years from now.”

At Christmas 2022, with the intention of announcing to his family that he was leaving healthcare work to devote himself fully to generative AI, Prado arranged a surprise. “I gave my family an AI-generated painting. It had a QR that took you to a website with a video narrating an AI-generated story read by an artificial voice. They were quite impressed.” Then came the discussion. Prado wanted to return to the family event company to implement everything he had learned, but they didn’t want him to leave the job he had worked so hard to get. In the end he managed to convince them, stayed with the family business for another year, and then set up his own. The small project is focused on offering customers different experiences based on AI. The end product is better, and he is able to get it done faster and for a lower price than his family’s company. “That’s what’s going to make the difference.”

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition