Artificial intelligence is reshaping work in software, data, finance and marketing, but the labour-market effects are not one-directional. Jobs more exposed to AI are seeing faster pay growth in job advertisements: Indeed Hiring Lab data shows median advertised salaries in these roles have risen 46% since 2021, compared with a 25% increase for jobs with lower AI exposure.
The figures suggest that AI is not simply reducing the number of jobs. It is also changing which skills employers are prepared to pay more for. Economist Jack Kennedy said AI is “reshaping what the market is willing to pay for,” rather than simply replacing workers with specialised expertise.
Software, data and finance lead pay gains
Indeed Hiring Lab classifies roles as highly AI-exposed when AI is likely to perform some of their tasks or materially alter how the work is carried out. The group includes software development, IT systems and technical support, data and analytics, marketing, banking and finance.
By 2026, median advertised pay in these roles was 46% higher than in 2021. By comparison, jobs with lower AI exposure—including nursing, personal care and home health care, food preparation and service, cleaning and sanitation, and production and manufacturing—recorded a 25% increase. That pace had also slowed from a year earlier.
The gap varies by career stage. Among entry-level roles, the difference in pay growth between highly exposed and less exposed jobs is limited. The divergence is more pronounced higher up the career ladder, suggesting that AI may be having a stronger effect on the skills mix, duties and pay-setting of experienced roles rather than producing an identical shift for all jobseekers.
Hiring demand is also recovering in exposed roles
Kennedy said hiring opportunities in highly AI-exposed occupations had generally recovered over the past year. The number of job postings has been rising steadily while pay has continued to increase, a pattern consistent with employers placing greater value on the skills required for those positions.
The US Bureau of Labor Statistics publishes 10-year employment projections for hundreds of occupations. In 2026, it also incorporated several studies on AI to classify how different occupations may be affected by the technology.
Among the 206 occupations rated as having “very high” AI exposure, many offer relatively high pay and are still expected to add jobs between 2025 and 2035. Software developers and management analysts are examples. Physician assistants and dental hygienists, meanwhile, combine relatively high pay with lower AI exposure and are also expected to see employment growth over the period.
Employment fears persist as usage expands
Public concern about AI’s effect on employment continues to rise. A Pew Research Center survey found that 71% of US adults believe AI will reduce the number of jobs the country needs over the next 20 years, up from 64% in 2024. The survey also found that people in countries with higher GDP per capita are generally more concerned that AI will eliminate jobs; in Australia and South Korea, the figure was 76% in each country.
Some views within the AI industry are more severe. Jacob Coxon, who previously worked at OpenAI and Anthropic, said on X this month that he had resigned from Anthropic and argued that people seriously developing AI believe the technology could cause the destruction of humanity before the end of this decade. The statement reflects his personal view, not an industry consensus, but it illustrates that debate over AI’s limits and potential consequences remains active.
Businesses and individuals, however, have not stopped using the technology. More workers are applying AI to automate repetitive tasks and speed up workflows, while others use it for advice in daily life. Meta’s personal AI agent recently reached the top of the US Apple App Store’s free-app rankings. Some users have also begun turning to AI chatbots for relationship and emotional support.
The data point to two parallel developments: pay and hiring demand are increasing in highly AI-exposed roles as the market reprices related skills, while concern about AI’s long-term employment effects is deepening. The eventual impact on individual occupations will depend on how companies reallocate tasks, how job numbers change and whether workers can acquire skills suited to the new tools.