Will we be servants of AI?

The trickle of bad news about the evolution of the Catalan labor market has set off some alarms. On the one hand, the EROs announced by Nissan, Nestlé, and Mediapro. On the other, the slowdown in job growth according to data from the Active Population Survey.

One of the causes, undoubtedly, is the geopolitical context. Although our attention, always volatile, has already turned to hantavirus and sporting joys, the war with Iran and the blocking of Hormuz, fully integrated into our daily lives, continue to affect inflation and harm growth prospects for this year. And, in the background, the rumblings of the impacts of generative artificial intelligence, which seems to be everywhere and nowhere at the same time. Nissan accompanies its ERO by signing a strategic agreement with a British AI company to incorporate it into its new vehicles. Nestlé justifies the reduction in personnel by alluding to digitalization and a new business model, and Mediapro operates in a sector in transformation. But these are generic and vague words. The impact of AI on the labor market will require much more specific indicators.

It must be said that it is not always easy to transfer this technology to the business world: lack of knowledge, failed tests, excessive cost, or imperfect agents. Therefore, equipping ourselves with the ability to distinguish between potential and reality is essential. In this regard, Anthropic proposes differentiating between the “theoretical capacity” of these models and the “real capacity”, that is, the one that is grounded in the day-to-day of jobs. According to a study by this AI company, some jobs with a very high margin of automation, over 90%, such as finance or data entry, currently incorporate it in only 33% of tasks. This is just the beginning. And what is currently observed in the labor market is a decrease in the hiring of young people, recent graduates, highly educated but without experience, in those jobs most susceptible to automation.

But this technology, which already accounts for 60% of global investment in start-ups, is improving exponentially, and its adoption rate, currently manageable, may cease to be so. Faced with the expectation of an AI that can improve itself, predicted between 2028 and 2030, I find the quote from Harvard student Chanden Climaco to be apt when he says that “we are creating machines that will be much more intelligent than us, we are doing it as quickly as we can, and we have no idea what will happen when they arrive”.

Meanwhile, for the futurists, take a look – but don't sign up – at the website RentAHuman, which connects people with AI agents that require a physical body to develop some of their functions: running errands, conducting visual checks of spaces or providing sensory descriptions, among others. 900,000 people have already signed up worldwide. I entered but did not complete the registration. I still have a job and my bosses are made of flesh and blood. For now.