The British computer scientist Geoffrey Hinton, a pioneer in the development of Artificial Intelligence, recently announced his decision to leave his job at Google to warn more freely about the dangers of this technology. In this sense, ‘the godfather of AI’ recognized in conversations with The New York Times that the rapid development of Artificial Intelligence systems was “scary”.
From the great potential of the IA. In addition, Hinton showed his concern for the possible impact of tools like ChatGPT in the labor market. The computer scientist, who received the Turing award in 2018 for his work in the field of neuronal networks, pointed out that this type of chatbots can remove the most tedious work, warning that, perhaps, they could “remove something more than that”.
Changes at a global level. In relation to this issue, Goldman Sachs published last March 26 a study about the possible alterations of AI in the workplace. The report indicates that, according to estimates based on data from the US and Europe, almost two thirds of current jobs were exposed to a certain degree of AI automation.
300 million workers replaced. Concretely, Goldman Sachs analyzes indicate that 300 million jobs worldwide could be completely automated. However, the bank indicates that the automation of jobs “has historically compensated for the creation of new jobs”, a process that favors “long-term job growth”.
economic growth In fact, the entity headed by David Solomon points out that the automation of jobs allows reducing labor costs and increasing productivity, allowing a “productive boom that increases economic growth substantially, although from the moment of said boom it is difficult to predict” . Additionally, the report estimates that AI could generate an annual increase of 7% of global GDP.
Automation in the office. On the other hand, the study points out that in Europe, exposure to AI automation is highest in jobs that are carried out in an office (45%), those that require a high level of academic qualification (34% ) and those linked to more technical tasks (31%). On the contrary, jobs linked to care (8%), production lines (7%) and manufacturing (4%) are the least exposed to said automation.
Most exposed countries In addition, Goldman Sachs points out that Hong Kong, Israel and Japan have the labor markets that are most exposed to AI automation, with a percentage higher than 25%. For its part, the exposure of the EU labor market is exactly in that percentage, higher than the global average of around 20%.
Less work thanks to AI. At the same time, researchers from OpenAI, OpenResearch and the University of Pennsylvania, prepared a report published last March 27 in which they analyzed the potential impact of LLMs (‘Large Language Models’ in English) in the labor market. According to research, 80% of jobs in the US would be exposed to the automation of 10% of their tasks. In this sense, 19% of American workers would see half of their work automated.
They differ socioeconomically. The document, made through the analysis of the GPT-4 application, points out that if automation is good, it is transversal in terms of wages, it is specifically higher in those jobs whose salaries are higher. In addition, the text indicates that the programs based on LLMs would accelerate by 15% all the tasks of the workers in the US with a quality similar to the current one.
The automation by sectors. Additionally, the study points out that some of the sectors with greater exposure to automation through AI, specifically from the GPT4 language model, are from data processing, book publishing, insurance companies, and scientists. Conversely, the sectors with a lower degree of exposure are the timber industry,
For example IBM. On the other hand, the rapid development of AI prevents researchers from determining an exact time in which the potential automation of jobs that, currently, are performed by people, will be completed. Sin embargo, it seems that a previous step is the freezing of contracts for jobs that, in the future, can be carried out for an IA. That was the decision taken recently by IBM: the company has decided to stop its recruitment process for positions that, theoretically, can be filled by an AI.
From replacement In this sense, the CEO of IBM, Arvind Krishna, affirmed that 30% of the company’s current jobs could be “replaced by AI and automation in a period of five years”, as reported by Bloomberg recently. This would mean, according to the North American publication, the elimination of 7,800 jobs.
A primer step. It seems to be the first part of the replacement of workers by the IA, which, on the other hand, is generating another type of jobs due to its rapid development. It will be expected to know the jobs that AI is capable of creating, specifically, in each of the sectors with greater exposure to automation.
The historical rhyme. Here it is interesting to remember a study published in August 2022 by the American National Bureau of Economic Research. According to the report, the majority of current workers in the US are employed in jobs that did not exist before 1940. The analysis also indicates that the appearance of new jobs “responds to technological innovations”.
Image: Safar Safarov / Unsplash
In Xataka | The EU already knows how to protect copyright against ChatGPT and other IA. Y quiere convertirlo en ley