The work that is gone and the work that will be

AI has become the first defendant in the current professional revolution. Everyone’s job is at risk, especially for those who have studied the most. But the impact is (and will be) so significant that it will generate new consumption and, therefore, new jobs

Lidia Baratta
Reading time: 4'45"
Credit: NG HAN GUAN / AP PHOTO


In 1983, Wassily Leontief, Nobel Prize for Economics, said that human labor would meet the same fate as horses after the arrival of the automobile: “First it is reduced, then it is eliminated.” More than forty years later, a new wave of catastrophic predictions revolves around artificial intelligence, the technology that has sparked fresh fears about the replacement of human labor by machines, fueling dystopias of a “world without work,” where robots produce everything and algorithms provide all necessary services.

The pattern has repeated itself many times throughout human history: industrial progress has always fascinated humans, but at the same time, it has also frightened them. The truth, however, is that in the relentless technological progress of the last eighty years, the labor market has transformed and re-transformed several times. And the same is likely to happen now. Because while it is true that some professions will become obsolete with AI (as has already happened in the past), others could benefit in terms of productivity, and still others will emerge. The central issue is understanding how artificial intelligence will integrate into the workforce. For example, the investment bank Goldman Sachs has predicted that 300 million jobs worldwide will be exposed to automation due to AI, but it clarified that the “integration” between algorithms and humans will be greater than the “replacement.” The study “Generative AI and Jobs: A global analysis of potential effects on job quantity and quality” by the International Labour Organization (ILO) further explains that generative AI is more likely to increase jobs rather than destroy them, by automating certain tasks and creating new professional opportunities.

Much will depend on the skills that the affected workers possess. According to a study by the AI Enabled ICT Workforce Consortium, which includes tech giants like Cisco, Google, Microsoft, and Intel, with the improvement of AI tools, “some skills will gain importance (such as ethics and AI literacy), while others may become less relevant (traditional data management, content creation, documentation maintenance, programming).” However, 87% of the managers interviewed expect job roles to be expanded, rather than replaced. What is urgently needed, the report explains, is to launch reskilling and upskilling initiatives for workers. Companies must take action, but public resources will also be necessary.

DATI: IA INDEX 2024 (STANFORD UNIVERSITY)

Most common features of AI in business


We are not yet able to properly measure the force of the AI tsunami on the business world, but many people have begun to do so. And, somewhat surprisingly, we are beginning to realize that catastrophism is not the right way to read this reality.
Personalization of services
Writing documents, notices, reports
Creating marketing campaigns

The impact of the arrival of AI on core business tasks.

A map

According to the “Artificial Intelligence Index Report 2024” from Stanford University, the two most common functions where AI is used in companies are the automation of contact centers with virtual assistants (26%) and service personalization (23%). Generative AI is primarily used by workers to draft initial versions of documents (9%), design personalized marketing campaigns (8%), summarize long texts (8%), and create images or videos (8%).

In Italy, where the AI market grew by 52% in 2023, reaching a value of 760 million euros, the most significant market share is linked to solutions for analyzing and extracting information from data (29%), according to the Artificial Intelligence Observatory of the School of Management at Politecnico di Milano. Another 27% is dedicated to projects for interpreting written or spoken language, 22% to algorithms that recommend content to customers based on their preferences, and 10% to video and image analysis.

In this context, the highest demand for jobs requiring AI-related skills in 2023 was recorded in the United States, Spain, and Sweden. However, the countries with the highest hiring rates for AI related professions were Hong Kong, Singapore, and Luxembourg. The most sought-after skill is machine learning management, followed by interaction with AI and Natural Language Processing (NLP). Stanford University’s AI Skill Penetration Rate indicates the intensity of demand for these skills across occupations worldwide. From 2015 to 2023, the countries with the highest AI skill penetration rates were India (2.8), the United States (2.2), and Germany (1.9). In the U.S., the penetration of these skills was 2.2 times higher than the global average in the same set of occupations. Italy ranks 13th in the top 15 with an average rate of 1.08%, but there is a significant gender gap, with 1.10% for men and only 0.46% for women.

A map - Credit: NG HAN GUAN / AP PHOTO

Above: Despite amazing advances in robotics, we are lagging behind in emulating more physical jobs.

On average, seven out of ten workers are exposed to the effects of AI. However, human oversight will always be needed in areas such as medicine and justice.

How AI is Growing

According to Stanford researchers, in 2023, artificial intelligence surpassed human performance in several areas, including image classification, visual reasoning, and basic text comprehension. However, it still lags far behind in more complex tasks, such as advanced mathematical problem solving, commonsense-based visual reasoning, and planning.

The area that has garnered the most public attention is the interpretation, comprehension, and generation of language, both written and spoken, which includes chatbots and Natural Language Processing (NLP). More advanced models, such as OpenAI’s ChatGPT-4 and Google’s Gemini, allow for the generation of fluent prose and high levels of language comprehension, even with inputs beyond written words, like images and audio.

Despite these advancements, large language models (LLMs), which power these systems, still frequently produce errors or false information. Therefore, as Stanford researchers suggest, human oversight remains essential, especially in fields like medicine and law, where these technologies are increasingly used to make critical decisions.

Focusing on Tasks

Focusing on Tasks - Credit: MATTEO SECCI / AP PHOTO

In the logistics industry, AI is already being used to optimize warehouse space management.


As the World Economic Forum explains, replacing workers with artificial intelligence is more difficult than it seems because “jobs consist of a set of tasks, and software may not be able to perform all of them seamlessly.”

The best analysis, therefore, focuses on the specific tasks that AI can handle, to understand which professions may be most affected by its introduction. A group of Italian researchers and economists— Guido Baronio, Antonio Dalla Zuanna, Davide Dottori, Elena Gentili, Giovanna Linfante, and Luca Mattei—focused precisely on the human skills used in various professions to determine the degree of exposure to AI.

The concept of “exposure,” they explain, does not imply replacement, but rather a general “interrelation” that may even lead to complementarity, with potential benefits in terms of increased worker productivity. Since AI is more connected to cognitive skills, the occupations most exposed are those where such skills are most used. For instance, a lawyer must be able to “organize information.” Because AI can interact significantly with this skill, lawyers are considered highly exposed to AI, at least for this task.

The researchers calculated that in Italy, over seven out of ten workers (slightly more than 15 million out of around 21.5 million) are in professions potentially a affected by the introduction of AI systems. For nearly 7 million, or one-third of the entire workforce, the exposure will be high. The innovative aspect of this technology is that it is the more educated workers who are most exposed to technological change. Those with a university degree are classified as middle- or high-exposed in 95% of cases. Particularly significant is the proportion of highly exposed workers in sectors like services, public administration, information and communication, finance, inurance, education, healthcare, and other social services. Since women are employed in these sectors 37% of the time compared to 17% for men, they are more exposed to these changes. However, it’s not only highly skilled professions that are more affected. High exposure is also seen in administrative jobs, characterized by high levels of “human-computer hybridization.” According to the International Labour Organization (ILO), the greatest impact is likely to occur in high- and upper-middle-income countries due to a larger share of administrative positions, which are the most exposed to AI integration.

This is not necessarily about replacement. The research notes that “the most exposed professions may be able to exploit complementarities to increase their productivity and, therefore, their wages.” For this reason, it’s important to monitor not only the risk of reduced hiring for the most affected professions but also income trends, particularly in the service sector. Much will depend on the kind of “coexistence” we can build with this innovation and the skills we can develop to make the most of these technologies, perhaps leaving automation to handle more elementary tasks while focusing on the creative aspects of our work. According to the ILO, AI could help us save time and further enhance the human contribution to work by reducing time spent on routine tasks and focusing workdays on higher value-added activities. Algorithms won’t make final decisions on contracts or court cases, but they can provide basic texts, quickly generating content from existing data. For this reason, jobs requiring critical thinking could become even more valuable.

How Is Coexistence Going

Studies reveal that AI is indeed helping workers across various sectors complete tasks faster and produce higher-quality work. A Microsoft study compared the performance of workers using Microsoft Copilot or GitHub Copilot with those who didn’t. It found that the former group completed tasks between 26% and 73% faster than their colleagues without access to AI.

Similarly, a study by Harvard Business School showed that consultants with access to ChatGPT-4 increased their productivity by 12.2%, their speed by 25.1%, and the quality of their work by 40%. Research from the National Bureau of Economic Research highlighted that call center agents using AI handled 14.2% more calls per hour than those who didn’t. The legal profession has seen similar improvements: AI’s impact on legal analysis demonstrated that teams using ChatGPT-4 significantly boosted efficiency and improved quality, especially in contract drafting.

One surprising aspect emerging from these studies is that access to AI appears to reduce the performance gap between less-skilled and highly skilled workers. Harvard’s study on consultancy tasks revealed that participants with lower educational qualifications showed a 43% improvement, while those with higher skills had a 16.5% increase. While more skilled workers using AI still outperformed their less skilled counterparts, the performance gap was significantly smaller when AI was utilized.

POLITECNICO DATA

AI market growth in Italy in 2023

How Is Coexistence Going

According to the Observatory of the School of Management of the Politecnico di Milano, in a market that grew by 52%, the most significant share is related to solutions for analyzing and extracting information from data. Other applications include: organization and storage management or personalized marketing.

Algorithms that suggest content to customers according to their preferences
How Is Coexistence Going
 
Language interpretation activities (written or spoken)
How Is Coexistence Going

Video and image analysis
How Is Coexistence Going

Work is Transforming, Not Disappearing

According to the World Economic Forum, it is important to view the changes from a long-term perspective. Philipp Carlsson-Szlezak and Paul Swartz argue that technology is an “intrinsically deflationary force.” When technologies like AI have a broad impact on our lives, they reduce costs and prices, increasing real consumer incomes and driving demand for new goods and services, which in turn creates new employment opportunities.

A clear example is the transformation in the food industry. At the end of the 19th century, nearly half of Americans worked on farms and spent over 40% of their income on food. Over the next 150 years, the introduction of machinery reduced this to about 1% of Americans working on farms today, while food expenditure has dropped to around 12% of income. Lower food prices led to higher real income. Consumers spent less on food and allocated the extra budget to new goods and services, creating new jobs.

It is true that the process initially involves job losses in some areas, but it is also accompanied by the creation of new jobs. The same will happen with artificial intelligence. According to the World Economic Forum’s white paper, “Jobs of Tomorrow: Large Language Models and Jobs,” which examined over 19,000 tasks across 867 occupations, the jobs most at risk of extinction include bank tellers, postal workers, cashiers, and data entry clerks.

However, new roles will emerge, such as AI and machine learning developers, business intelligence analysts, interface designers, and specialists in AI ethics and governance. Jobs in education, career counseling, and training will remain relatively unchanged.

It seems unlikely that AI will end a history of labor market renewal and adjustment,” the WEF explains.

Therefore, it is improbable that humans will face the “fate of the horse” predicted by Leontief. A much more likely outcome is a gradual increase in productivity and wealth, accompanied by job losses, relocations, and skill retraining that have characterized previous economic transformations.

AI could also address the challenge of an ageing population. As companies struggle to find employees, the demand for technologies to address staff shortages and improve the productivity of older workers will rise. The world won’t be without work, but it will operate differently.

 

Work is Transforming, Not Disappearing

Lidia Baratta
Business editor at Linkiesta.it, she writes a weekly newsletter called Forzalavoro. She has collaborated with D di Repubblica, L’Espresso, La Stampa and Vice and is one of the radio hosts of Prima Pagina (RadioTre).