# AI Is Destroying the Job Market — So Why Are New Jobs Appearing? > Source: > Published: 2026-08-12 13:38:37+00:00 The job market is sending two messages at once. Technology companies are hiring fewer junior workers. Customer service and administrative jobs face growing automation pressure. Employers are openly discussing headcount reductions. At the same time, job postings that mention AI are rising. Data scientists, AI engineers, cybersecurity specialists, and software developers remain among the fastest-growing occupations. Both trends are real. AI has not caused a large collapse in total employment so far. The clearest change is happening inside occupations. Employers are reducing demand for some tasks while hiring people who can build, manage, or work with AI systems. The total number of jobs may continue growing while access to those jobs becomes harder for people without the newly required skills. *The figures below come from several datasets. Indeed, Lightcast, the US Bureau of Labor Statistics, and the World Economic Forum measure different things, so their percentages should not be compared as if they came from one continuous survey.* There was no clean “before AI” period. Companies were already hiring machine-learning engineers, data scientists, computer-vision specialists, and robotics engineers long before ChatGPT appeared. The more useful dividing line is November 2022, when ChatGPT brought generative AI into everyday work. Indeed’s data shows that AI-related hiring had already reached a high point before that launch. In March 2022, 3.3 percent of US job postings on Indeed contained AI-related terms. Those postings were heavily concentrated in software development and mathematics. This was also a period of unusually strong technology hiring. Companies expanded during the pandemic, digital services grew quickly, and employers competed for software and data workers. When the technology market cooled, AI hiring initially fell with it. That matters because some of the later decline in tech employment cannot be attributed to generative AI alone. Higher interest rates, post-pandemic overhiring, slower corporate spending, and layoffs at large technology companies were already reshaping the market. The first year after ChatGPT produced an unusual split. General technology hiring fell, but demand for generative AI knowledge increased. Indeed found that the share of US job postings mentioning AI dropped from 3.3 percent in March 2022 to 1.64 percent in June 2023. It then began recovering. Generative AI jobs were still rare, but their growth was fast. In January 2023, only about 3 in every 100,000 US job postings on Indeed mentioned generative AI. By the end of February 2024, that figure had reached 11 in every 10,000 postings. That is more than a 30-fold increase in roughly 14 months. Lightcast recorded a similar change. Its data, used in the Stanford AI Index, found that: General tech hiring remained weak during this increase. By February 2024, software development postings on Indeed were 67 percent below their March 2022 level. Yet 22 percent of the software development jobs still being advertised mentioned AI. The market did not return to its previous size. Employers changed the skills they requested within a smaller pool of openings. By the end of 2025, AI hiring had moved further away from the rest of the labor market. According to Indeed Hiring Lab, 4.2 percent of US job postings mentioned AI in December 2025, the highest share recorded by its tracker. The number of all job postings was only 6 percent above its February 2020 baseline. Postings that mentioned AI were 134 percent above that baseline. The difference was particularly large in technology: AI requirements also moved into nontechnical occupations. In marketing, the share of postings mentioning AI rose from 8.4 percent at the beginning of 2025 to 14.9 percent in December. In human resources, the share doubled from 4.4 percent to 8.8 percent. This does not mean every marketing or HR worker has become an AI specialist. Employers are adding AI-related tasks to existing jobs. The title may stay the same while the hiring standard changes. The strongest warning appears in entry-level hiring. A Stanford study using payroll records from 25 million US workers found that employment among workers aged 22 to 25 declined 13 percent in occupations with the highest AI exposure after 2022. The Federal Reserve Bank of Dallas reviewed that result using Current Population Survey data. Its analysis also found a decline among young workers in highly exposed occupations. The share of young workers employed in those occupations fell from 16.4 percent in November 2022 to 15.5 percent in September 2025. The decline did not appear to come mainly from layoffs. Young workers were having more trouble entering highly exposed occupations in the first place. The job-finding rate for young labor-market entrants seeking highly exposed work fell by more than 3 percentage points from its recent peak. The researchers warned that the pattern may not be causal. AI-exposed occupations differ from other occupations in education, industry, and sensitivity to the business cycle. The aggregate effect was also small. The Dallas Fed estimated that even if the entire decline translated into unemployment, it would explain only around 0.1 percentage point of the increase in the overall unemployment rate since November 2022. The measured problem is concentrated. AI has not produced mass unemployment, but some young workers are finding fewer entry points into occupations where junior tasks can be automated. The broader data does not support the claim that AI has already destroyed the US job market. A Federal Reserve analysis examined Lightcast job postings and Census Bureau survey data from more than one million firms. The researchers found no evidence that industries or companies with higher AI adoption were posting fewer jobs. At the firm level, the estimated effect was close to zero. Companies requesting AI skills did not show a later collapse in total postings. The Yale Budget Lab reached a similar result. Its analysis found no statistically clear effect of AI exposure on employment or inflation-adjusted hourly wages. The US labor market has weakened, but AI is only one possible factor. Interest rates, slower economic growth, post-pandemic normalization, and reduced worker turnover also affect hiring. The current evidence supports a change in hiring priorities more strongly than a collapse in total employment. Some new jobs involve building AI directly: Other jobs combine AI with an existing field: Lightcast analyzed more than 1.3 billion job postings and reported that 51 percent of postings requesting AI skills were outside IT and computer science in 2024. The same analysis found that job advertisements requesting AI skills offered salaries 28 percent higher on average, or nearly $18,000 more per year. That figure is an advertised-pay difference, not proof that learning AI automatically raises someone’s salary by 28 percent. AI jobs may already be concentrated in higher-paying occupations, industries, and locations. It does show where employers are placing value. They increasingly want workers who understand a professional field and can apply AI within it. The data points to three hiring changes. Junior employees have traditionally learned through tasks such as research, basic coding, document review, data entry, and first-draft preparation. Generative AI can now complete parts of that work. A company may still hire senior workers to review output, manage clients, and make decisions while reducing the number of junior employees needed to produce the first draft. This helps explain why entry-level workers can struggle even when experienced employment remains stable. A marketing position may now require experience with AI content tools. A financial role may request machine-learning knowledge. A software role may require experience integrating language models. The company is not always creating a separate AI department. It is adding AI expectations to existing positions. If an employee can use AI to draft reports, analyze information, and automate routine work, the employer may combine responsibilities that were previously divided among several people. This does not always remove a job immediately. It can reduce the number of new workers hired as the company grows. That effect appears in hiring data before it appears in layoff announcements. The World Economic Forum’s Future of Jobs Report surveyed more than 1,000 employers representing over 14 million workers in 55 economies. Across technology, demographics, trade, economic conditions, and the energy transition, employers projected: These numbers cover several economic forces. They are not an AI-only forecast. The report separately estimates that AI and data-processing technologies could create 11 million roles while replacing 9 million by 2030. Among surveyed employers: The fastest-growing roles by percentage included big data specialists, fintech engineers, and AI and machine-learning specialists. The largest job gains by total number were expected in frontline and essential work, including agriculture, delivery, construction, care, and education. New employment will not consist entirely of AI engineers. The US Bureau of Labor Statistics expects AI adoption to increase demand in some technical occupations while reducing demand in administrative and repetitive information-processing work. Its 2024 to 2034 projections include: These projections include more than AI. Industry growth, consumer demand, demographics, and other technologies also affect employment. The direction is still clear. Jobs involving data, software, security, and technical decision-making are expected to grow. Jobs centered on routine information handling face weaker demand. The available evidence does not show an AI-caused collapse in total employment. It shows a narrower route into many knowledge-work careers. Employers are hiring fewer generalists in weak sectors while increasing demand for workers who can use AI inside a specific occupation. Entry-level candidates face the most immediate pressure because many of their traditional training tasks are easier to automate. The strongest position is domain knowledge combined with practical AI ability. That might mean a marketer who can measure AI-assisted campaigns, a financial analyst who can test model output, or a developer who can integrate an AI system and verify its behavior. Prompting alone is unlikely to be enough. Employers are looking for evidence that a candidate can use AI to complete real work accurately, safely, and with less supervision. New jobs are appearing because AI creates demand for technical specialists and changes existing positions. Other jobs decline because the same technology reduces the amount of labor required for routine tasks. The job market is not simply losing jobs or gaining jobs. It is replacing one set of hiring requirements with another. [AI Is Destroying the Job Market — So Why Are New Jobs Appearing?](https://blog.stackademic.com/ai-is-destroying-the-job-market-so-why-are-new-jobs-appearing-134bdca87af0) was originally published in [Stackademic](https://blog.stackademic.com) on Medium, where people are continuing the conversation by highlighting and responding to this story.