[2/2] Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence (AI column)

Continue. story About canaries, it is necessary to tell that the methodology was based on ADP data on salaries, as well as on the assessment of the susceptibility of professions to the influence of AI, which used the approach from the ADP.GPTs are GPTs: Labor market impact potential of LLMs" and the Anthropic Economic Index" ("Which Economic Tasks are Performed with AI? Evidence from Millions of Claude Conversations"). Anthropic’s Economic Index is Based on Analysis of the Real Use of AI (Claude) Millions of queries, showing how much of the queries are related to specific occupations. It allows you to divide the applications of AI into automation vs. complementary: the index estimates how many requests are automatic. (completeness) and how much is auxiliary (helper). This allowed the authors of the canary study to classify occupations by whether AI is used as a substitute for labor or as a tool to improve labor efficiency.

By comparing ADP personnel data with exposure figures, the researchers tracked employment dynamics since the end of the year. 2022 midway 2025 years. The starting point is autumn 2022When there was a jump in the availability of generative AI tools (For example, the appearance of GPT-3.5/ChatGPT). The analysis was carried out by groups: they compared professions with high and low exposure to AI, and within them – young specialists vs. older employees. To isolate the effect of AI, the authors applied firm-level controls. (Take into account shocks and trends within individual companies and industries)

In the future, the authors emphasize not only the short-term impact on employment, but also the adaptive mechanisms. Historically, technological revolutions (computerization 1980-90.) This was followed by a period of redistribution of jobs, followed by new growth in employment and wages. Perhaps a similar process is unfolding now: there are signs that young people are already reacting to the trend, reorienting towards less “automated” professions. (Decreased interest in specialties closely related to AI, such as computer science).

Industries will have to adapt: companies that are actively implementing AI must think through strategies to retrain staff and create new roles where human labor complements AI. In the short term, individual sectors of the economy will have to address the problem of youth employment – to avoid the “lost generation” of novice professionals displaced by algorithms. In the long run, if adaptation succeeds, AI could become a tool that will increase productivity and open up new opportunities for growth in most industries, as has happened with previous waves of technology. But the current study makes it clear that AI’s impact on the labor market has already begun, and industries should not ignore its signals.

#AI #Software #Engineering #Management #Leadership #Data #Whitepaper #Metrics