AI job – New data points to a workforce shift that’s no longer hypothetical: employment for 22–25-year-olds in high-AI-exposure jobs fell 6% between late 2022 and July 2025, even as AI-related hiring expands. The fallout is uneven—women, entry-level roles, and customer
For a worker at the start of their career, the change doesn’t arrive like a dramatic layoff. It shows up sooner—fewer callbacks, tighter entry pipelines, and a growing sense that “how work gets done” has quietly moved on without asking whether you were ready.
The numbers now put a date on that feeling. Employment for 22–25-year-olds in high-AI-exposure jobs fell 6% between late 2022 and July 2025. while employment for workers aged 30 and older in the same jobs grew by 6–13%. The gap is stark enough to matter for families making near-term decisions about training, careers, and stability.
Even broader displacement concerns are also meeting a counterweight: organizations are increasingly comfortable handing decisions to AI in low-risk areas. Nearly 50% of buyers say they would grant full autonomy to AI agents in low-risk workflows. signaling a willingness to move beyond “assist me” toward “decide for me.”.
This combination—less room at the entry level, more autonomy at the workflow level—helps explain why job displacement fears are becoming harder to dismiss as something that will come later.
AI at work has expanded from curiosity to routine. driven largely by the rapid adoption of generative AI tools that take over tasks that once required time and manual effort. Drafting content. handling customer queries. and assisting with code are among the everyday areas where AI is moving from toolbelt to backbone.
Software is reflecting that shift. In early 2026 alone, G2 introduced 11 new software categories, including 5 focused entirely on AI. Those categories also included niche additions like AI interview agents and AI medical diagnostic platforms—an indicator that AI is no longer a sidecar feature. but a core part of workflows across industries.
At the same time, the displacement signal is showing up differently depending on where you look. Administrative and data entry roles are projected to decline by 3.9% by 2034. with data entry. documentation. and rule-based admin tasks singled out as highly automatable. In workflow software reviews. 12% of G2 reviews mention reduced manual work. describing how tasks once split across a team are now handled by a single person with the right tools.
Customer support is a similar story, but with different stakes. 70.1% of customer service tasks show AI exposure. and basic support and repetitive queries are increasingly handled by chatbots and AI agents. In G2 review patterns. 60% of reviews in the customer success category mention AI or automation. reflecting how central these tools have become to day-to-day support operations.
Manufacturing and operations carry their own warning. The human task share in manufacturing is expected to drop from 43% to 31% by 2030. Routine operational processes are gradually shifting to machine-led execution. In G2 manufacturing intelligence tools. 12% of reviews cite AI-driven efficiency gains. and the reporting links efficiency gains directly to headcount decisions.
Human resources is both target and accelerator. 19% of HR teams report job loss due to AI. Resume screening, interview scheduling, and coordination tasks are becoming less manual, while 47% of G2 reviews in the AI agents for HR category report positive workflow impact.
Sales, often treated as a relationship-driven field, also faces automation pressure. 10% of sales teams report job loss due to AI, with AI automating parts of prospecting and outreach workflows. In G2 reviews for AI sales assistant tools, 40% cite efficiency gains.
Marketing and content show a different kind of displacement dynamic—less about replacing every role overnight and more about squeezing volume and entry-level output. 73% of marketers say their budgets are more heavily scrutinized than they used to be. Initial content drafting, basic copywriting, and repetitive campaign tasks are being handled by AI tools. In marketing automation reviews, 26% mention AI handling content or campaign execution.
Software development and IT are the most “AI-saturated” section of this story. 42% of software development tasks can be automated. though a similar share is more likely to be AI-assisted rather than fully replaced. In G2’s AI IT agent tools category, 8% of reviews reference AI assisting with coding or development workflows.
The market signal is not just about tasks; it’s also about hiring patterns and timing. One set of figures points to the immediate pullback from the most automation-prone work. Postings for the most automation-prone occupations fell 17% after generative AI launched, while postings for augmentation-prone roles increased 22%. At the global level, AI job displacement vs. creation statistics still suggest more roles are coming than disappearing: 170 million new roles projected versus 92 million displaced by 2030. a net gain of 78 million. The catch is that the jobs created and the jobs displaced are not the same jobs—different skills. different wages. and different places.
That “transition gap” becomes sharper in a U.S. snapshot. In Q1 2025, the U.S. had 35,445 AI-related roles, a rise of 25.2% from Q1 2024 and an 8.8% increase compared with the previous quarter. Even so. only 6% of workers believe AI use in the workplace will result in more job opportunities for them over time. while roughly 32% think it will decrease their opportunities.
The distribution of risk is also different across demographics. The workers most affected by AI-driven job displacement are described as younger (22–30). more likely to be women. and concentrated in administrative. clerical. and customer-facing roles in high-income economies. Workers with AI skills are described as comparatively insulated, earning significantly more.
Women. in particular. appear to be pushed into the roles where automation is arriving first. while being underrepresented in the AI-specialist roles that are growing. Women are more likely to hold roles that are being automated and less likely to hold AI roles that are being created. A compounding-disadvantage dynamic is highlighted through research describing more displacement exposure. less augmentation benefit. and less access to new high-growth AI roles.
The numbers sharpen that point: women make up just one-third of the workforce in technology, information, and media roles globally. Worldwide, 4.7% of jobs held by women are classified as the highest risk, compared with 2.4% of men’s jobs. Among 22–25-year-olds. employment in high-AI-exposure jobs fell 6% between late 2022 and July 2025. while workers aged 30 and older in the same jobs grew by 6–13%. Despite that, 46% of Gen Z report worry about AI’s impact on their careers.
For employers, the planning shift is already visible. 23% of organizations anticipate direct job displacement in the near term. 40% of employers plan to reduce headcount in roles where AI can do the work.
In some functions. AI is also compressing workload so quickly that hiring needs don’t look like “full replacement”—they look like fewer new openings. Among applicants. an AI-led initial screening trial is described as nearly halving recruiter workload. reducing the number of human interviews required by 44%.
Confidence among software workers may also be changing. 92% of U.S. developers are reported to be using AI coding tools both professionally and personally. making the occupation the highest task coverage of any occupation in Anthropic’s framework. Yet developers’ confidence is described as waning: 64% view AI as not a threat, down from 68% last year.
Still, the upside of building AI skills appears measurable. Workers with AI skills earn 56% more than identical peers without them, and that premium doubled in a single year.
The picture that emerges is less about an even race between “jobs lost” and “jobs gained. ” and more about who gets the bridge and who gets the gap. One set of broader estimates puts expanding current AI applications across the economy at risk of displacement for only 2.5% of U.S. jobs. but the same reporting emphasizes that AI-related job postings are roughly twice as likely to include parental leave benefits and three times as likely to offer remote work—changes that point to better working conditions for many of the roles being created.
Even so, there’s already a clear early signal of what is harder for newcomers: their window is narrowing. When employment for 22–25-year-olds falls while employment for older workers rises. the labor market is sending a message about who gets the benefit of AI adoption and who pays the entry-level cost.
For many workers, the question isn’t whether AI will affect their field. It’s whether they will be the person who can operate inside the new workflow—especially in industries where the market is bifurcating between augmentation-prone demand and automation-prone contraction.
And in the space between those two poles—between a 17% decline in postings for automation-prone occupations and a 22% increase for augmentation-prone roles—work is being reshaped in real time, not at some distant horizon.
AI job displacement AI agents generative AI workforce trends customer support automation HR AI screening sales automation marketing AI content software development AI employment 22-25 women and AI risk G2 AI Agents Insights Report