How does AI affect jobs and wages?

Numbers match the citations in the article. A working link isn't proof that a page supports a claim; check the quoted passage and date.

1. 1

   Follow the money: A startup-based measure of AI exposure across occupations, industries, and regions.

   PNAS nexus (Fenoaltea et al.)Published Jun 23, 2026Checked Oct 3, 2026

   “Our findings indicate that even though white-collar high-skilled occupations are theoretically highly exposed, they are heterogeneously targeted by AI startups. Roles involving routine organizational tasks, such as data analysis and office management, show significant exposure, while occupations involving tasks that are tied to ethical or high-stakes considerations-such as judges or surgeons-present lower AISE scores, despite technical feasibility for automation. Our approach challenges the conventional assumption that high-skilled jobs uniformly face high AI risks, highlighting instead societal desirability and market-oriented choices as critical determinants of AI exposure. Contrary to fears of widespread job displacement, our findings suggest that AI adoption will be gradual and shaped by social factors as much as the technical feasibility of AI applications. This framework provides a forward-looking tool for policymakers to monitor the evolving impact of AI and navigate a fast changing labor market landscape.”

2. 2

   AI exposure, occupational mobility, and post-transition job quality in China: labor-saving and labor-augmenting channels.

   Frontiers in psychology (Zhang & Liu)Published Jul 28, 2026Checked Oct 3, 2026

   “Second, among switchers, labor-augmenting exposure exhibits positive but attenuated origin-destination sorting in all three switching samples. Labor-saving sorting is positive in the minor-group sample, statistically indistinguishable from zero in the intermediate-group sample, and negative in the major-group sample. Third, among switchers, greater labor-saving exposure in the origin occupation is associated with longer hours in the minor- and intermediate-group samples and with lower skill-match satisfaction in all three samples, whereas favorable associations with origin labor-augmenting exposure are confined to specific samples. The study extends organizational-psychology research by linking instrumented AI exposure to occupational switching as an observable career-adaptation behavior and to the fit and demand conditions of destination jobs. The findings suggest targeted reskilling and organizational support for less-educated workers in occupations with high labor-saving exposure.”

3. 3

   Large language model exposure and precarious occupations: Unpacking relationships in the Canadian labor force.

   Scandinavian journal of work, environment & health (Jetha et al.)Published Jun 3, 2026Checked Oct 4, 2026

   “Four multivariate linear regression models with cluster-robust standard errors estimated the associations between LLM exposure and each dimension of precarity. A fifth multivariate model examined the relationship between LLM exposure and the multidimensional precarity index. Utilizing model coefficients, mean estimates of occupational LLM exposure were produced.ResultsUsing the multidimensional precarity index, our analysis showed that occupations characterized by low exposure to precarity had a significantly higher mean LLM exposure [mean 0.386, 95% confidence interval (CI) 0.356-0.417] compared to occupations with medium (mean 0.258, 95% CI 0.221-0.295), high (mean 0.260, 95% CI 0.194-0.328) or very high precarity (mean 0.205, 95% CI 0.136-0.275). Apart from earning adequacy, LLM exposure was also lower among occupations using each separate dimension of precarity.ConclusionOccupations most likely to be exposed to LLM are those where precariousness is lowest. These occupations have previously been sheltered from technological change. There is a need of examine the impacts of LLM on workers in job where the technology is prominent.”

4. 4

   Between Humans and Machines: The Impact of Robot Adoption on Workers' Health in China.

   Journal of occupational and environmental medicine (Yuan)Published May 14, 2026Checked Oct 4, 2026

   “ObjectiveTo examine how robot adoption affects workers' health in China.MethodsUsing China family panel studies data matched with regional robot penetration measures, we estimate health effects with regression models and explore heterogeneity and mechanisms.ResultsRobot adoption significantly reduces perceived, diagnosed, and mental health issues among manufacturing workers. In contrast, effects in nonmanufacturing sectors differ. Health impacts vary across worker groups, reflecting heterogeneous exposure and labor-market adjustment.ConclusionsRobot adoption may generate adverse health effects through labor displacement and reduced bargaining power. Strengthening social protection and skill training is essential to mitigate health risks associated with technological change.”

5. 5

   Invulnerability bias in perceptions of artificial intelligence's future impact on employment.

   Scientific reports (Barrera-Jimenez et al.)Published Aug 6, 2025Checked Oct 3, 2026

   “The study analyzes survey data collected from 201 participants, recruited through social media using convenience sampling. The data were analyzed using a combination of statistical and machine learning methods, including the Wilcoxon test, ordinary least squares regression, clustering, random forests, and decision trees. Results confirm a significant IB, but not OBTI; only 31.8% perceived AI's future impact on their own job as more positive than on others'. Analysis shows that greater knowledge of AI correlates with lower IB, suggesting that familiarity with AI reduces the tendency to externalize perceived risk. Furthermore, bias levels vary across professional sectors: healthcare, law, and public administration exhibit the highest IB, while technology-related professions show lower levels. These findings highlight the need for interventions to improve workers' awareness of AI's potential future impact on employment.”

6. 6

   Explaining women's skepticism toward artificial intelligence: The role of risk orientation and risk exposure.

   PNAS nexus (Borwein et al.)Published Jan 20, 2026Checked Oct 3, 2026

   “Using original survey data from ∼ 3,000 respondents across Canada and the United States, we find that women consistently perceive AI to be riskier than men. We identify two key drivers behind this gender gap: women's higher general risk aversion and their greater exposure to AI-related risks. To establish a causal relationship between risk and AI attitudes, we show experimentally that as the probability of net positive employment effects decreases, women's support for companies adopting AI falls more sharply than men's. Finally, structural topic modeling of open-ended responses confirms that women express greater uncertainty about AI's benefits and more frequently anticipate little to no benefits. Given AI's potential to exacerbate existing gender inequalities, our study highlights the critical importance of incorporating women's perspectives into AI policy-making. Policies that do not address gender-specific risks may not only reinforce existing inequalities in employment and income but could also generate political backlash against AI adoption.”

7. 7

   Too old, too foreign, too replaceable? How AI shapes livelihood security of ageing immigrants.

   Frontiers in sociology (Mihajlov & Pavleska)Published Aug 4, 2026Checked Oct 4, 2026

   “A further finding, the foreignness premium, describes how AI systems expropriate the distinctive multicultural and regional professional assets that migration history conferred, turning these assets into drivers of late-career obsolescence. Additional themes document enforced professional silence near the retirement threshold, fragmented cross-border pension entitlements as a structural amplifier of displacement, and an identifiable set of buffer conditions under which Double Displacement does not fully materialise.DiscussionContextualised within the WHO Active Ageing framework, cognitive automation simultaneously undermines income security, social participation, and psychosocial health among this population. The findings call for age-decoupled reskilling provision, anonymous organisational reporting channels for AI-related dissent, and cross-border pension coordination to address the specific vulnerabilities of internationally mobile professionals approaching retirement.”

8. 8

   Algorithmic governance at work: a legal-sociological and policy analysis of AI regulation, institutional power and labor rights.

   Frontiers in sociology (Olimid et al.)Published Aug 19, 2026Checked Oct 3, 2026

   “The results indicate that AI regulation across Europe, the Americas, and the Asia-Pacific is increasingly intersecting around human rights and personal data protection, risk-based governance, and accountability mechanisms, particularly in the employment environment. Legal frameworks emphasize transparency, non-discrimination, and the need for human oversight in AI-powered workplace decisions.ConclusionThe conclusions highlight that across Europe, the Americas and the Asia-Pacific, AI regulation is increasingly organized around six dimensions: protection of human rights and data privacy in the workplace; governance frameworks addressing reskilling, job transformation and adaptation to AI-driven work models; risk-based oversight of AI systems; AI civil liability regimes affecting employee rights and HR practices; promotion of ethical AI to foster trust and mitigate bias; and strengthened monitoring, accountability and non-discrimination mechanisms.”

9. 9

   Artificial intelligence (Wikipedia)

   WikipediaPublished Oct 3, 2026Checked Oct 3, 2026

   “Artificial intelligence (AI) is the capability of computational systems to perform tasks typically associated with human intelligence, such as learning, reasoning, problem-solving, perception, and decision-making. It is a field of research in engineering, mathematics, and computer science that develops and studies methods and software that enable machines to perceive their environment and use learning and intelligence to take actions that maximise their chances of achieving defined goals. High-profile applications of AI include advanced web search engines, chatbots, virtual assistants, autonomous vehicles, play and analysis in strategy games (e.g., chess and Go), and content generation (e.g., text, images, audio, and videos). The traditional goals of AI research include learning, reasoning, knowledge representation, planning, natural language processing, and perception, as well as support for robotics. To reach these goals, AI researchers use techniques including state space search and mathematical optimisation, formal logic, artificial neural networks, and methods based on statistics, operations research, and economics.”