In November 2022, generative AI burst into our lives with the launch of OperAI's ChatGPT. Much of the debate since then has focused on how broad language models (LLMs) will affect the world of work, even in a negative sense (they are artificial intelligence models that use machine learning techniques to understand and generate human language). Even though we are talking about innovations, which have not yet developed their maximum potential, many workers around the world are particularly nervous. In the last year, according to Google Trends data, searches for “is my job safe”, associated with artificial intelligence, have grown by 180%. According to the latest report from the World Economic Forum: “Jobs of Tomorrow: Large Language Models and Jobs”, the work activities that will be hit hardest by LLM are those that use routine and repetitive language. At the same time, LLMs could further develop all those jobs that rely heavily on reasoning and problem-solving skills. In terms of numbers, 40% of jobs will be impacted by artificial intelligence. On the other hand, it is also correct to underline that alongside those who will lose or see their work modified/enhanced, new professions will also be born, such as artificial intelligence and machine learning specialists who will grow by an estimated 39% in the next five years.
Work and artificial intelligence
According to the report, there will be jobs such as software developers that will be replaced more easily than a human resources specialist. In detail, the work of a software developer can be divided into three macro bands: 28.7% of the work has a greater potential for automation, 43% has areas that can be enhanced by artificial intelligence and only for 28% are less susceptible to AI meddling. On the other hand, a human resources specialist sees only 16% automation in her work, compared to a 62% total inability on the part of the AI to find support points. This makes the work of human resources “safer” than that of software developers.
Jobs that will be replaced by artificial intelligence
As we have said, the jobs with the highest potential for task automation are those that emphasize routine and repetitive procedures and do not require a high degree of interpersonal communication. According to the research, the roles that present these characteristics are: office workers (81% of their work can be automated), management analysts (70%), telemarketers (68%), statistical assistants (61%) and cashiers (60%). Jobs with high potential for automation often include various types of office work, especially those focused on recordkeeping and information management. LLM models have demonstrated a strong degree of competence on these tasks. For example, legal secretaries and administrative assistants spend approximately 54% of their time on tasks with high potential for automation.
Jobs that will be enhanced by artificial intelligence
Then there are all those jobs where artificial intelligence can support. These jobs are mostly found in the fields of science, technology, engineering and mathematics. At the top of the list are insurance adjusters. According to the report, these workers dedicate 100% of their time to activities that have the potential to be enhanced by generative artificial intelligence systems. This is followed by bioengineers and biomedical engineers (84% of the increaseable time), mathematicians (80%) and editors (72%). The remaining 15 “upgradeable” jobs are always technical or highly specialized and often require advanced degrees or training, such as database architects and statisticians. Note that many jobs with the highest potential for improvement also have some significant level of automation, resulting in a very high exposure to being replaced entirely by AI.
The jobs least affected by artificial intelligence
Those least exposed to the arrival of artificial intelligence are all those jobs that require a high degree of personal interaction, such as healthcare workers or teachers, or physical movement, such as athletes. The job with the highest percentage of tasks classified as having low transformation potential, equal to 84%, is that of school, guidance and career counselors. This is followed by clergy (84%), home health aides (75%) and then anesthesiologists (74%). Professions linked to the community, social services and the health sector are therefore safer.
The off the charts
Beyond jobs with low exposure to automation, there are a number of professions that do not involve any linguistic tasks and will not be affected by LLM AI models at all. These jobs include: dishwashers, highway maintenance workers, carpenters, those who work in the textile world, those who are machine operators for paper products and a few others such as laminators.
