Employment in an Automated Era: October 2025 EN:Insights Forum

Are AI and automation destined to replace human labor? The latest SHRM research shows a more nuanced picture: Most workers aren’t in imminent danger of being displaced. But HR leaders can’t become complacent. They must continue to implement artificial intelligence strategies, address workforce anxiety, and help employees develop new skills for tomorrow’s jobs.

The October EN:Insights Forum, “Automation, Generative AI, and Job Displacement Risk in U.S. Employment,” examined new SHRM research based on the SHRM 2025 AI/Automation Survey  responses from more than 20,000 U.S. workers. The research explores how automation and generative AI (GenAI) are reshaping job tasks, how nontechnical barriers are blunting displacement, and which occupations are most exposed to future disruption.

The forum featured Justin Ladner, senior labor economist at SHRM, who shed light on where automation is advancing fastest, which sectors remain most resilient, and why leading organizations are focused on transformation — not on the possibility of AI taking jobs.

Below are four key insights from the presentation.

Research Insight No. 1: Most U.S. workers are unlikely to be displaced by automation in the near future, though even the small percentage of jobs that are highly exposed to automation could translate to millions of displaced workers.

Some occupational groups have become “highly automated” — meaning that at least 50% of tasks are affected by automation. But only 15.1% of U.S.  employment today fits that definition. Furthermore, a significant majority (63.3%) of these jobs include at least one nontechnical barrier to automation displacement (e.g., client preferences).

The SHRM research indicates that the vast majority of U.S. employment is plausibly shielded from automation displacement in the near term. Here are some key data points to know: 

• 94% of U.S. employment (about 145 million jobs) is either not currently highly automated or includes at least one nontechnical barrier to automation displacement (or both).
  • 54.2% of employment is both not highly automated and includes at least one nontechnical barrier to displacement.
• The remaining 6% of wage/salary employment (9.2 million jobs) will likely face elevated displacement risk because it is currently both highly automated and lacking in nontechnical barriers to displacement.
  • Although small in percentage terms, we could still see millions of workers displaced.

As of now, it seems likely that advances in automation and AI will mostly be geared toward job transformation, rather than displacement, Ladner noted. The goal of the research was to uncover “how automation technologies, in general, and AI, in particular, might influence either transforming or displacing work.”

Research Insight No. 2: Even in jobs that are currently highly automated, nontechnical barriers are likely to significantly forestall automation-based job displacement, at least in the near term. These include legal and regulatory limits, client preferences, and cost-effectiveness concerns.

For example, much of commercial flight is automated, but regulations and customer preference dictate that human pilots be involved. Small retailers, meanwhile, might forgo readily available automation because it isn’t cost-effective. 

The research further showed:

• 63.3% of jobs include at least one nontechnical barrier to automation displacement.
• 73.6% of jobs with at least one such barrier include a nontechnical barrier related to client preferences.
• 70.6% of employment in the health care practitioners’ occupational group has at least one nontechnical barrier to automation displacement, the highes among all major civilian occupational groups.

Clients and customers often care “that there's a human that's involved, that they're interacting with the human, that they have that level of trust and that level of interaction,” Ladner said.

Research Insight No. 3: Automation exposure varies sharply across occupations, with technical and production roles among the most exposed.

Automation is advancing fastest in computer and mathematical jobs, where GenAI tools have streamlined work processes. Production, architecture/engineering, and business and financial occupations also show above-average automation exposure.

Research findings include:

• 32% of computer and mathematical jobs meet the “high automation” threshold — the highest of any occupational group.
• 7.3% of education and library jobs are highly automated, the lowest of any occupational group.

Occupational groups with less automation are typically those that emphasize in-person interaction, Ladner said. “They also don't rely very much on the types of tools that have been emerging in recent years.”

Research Insight No. 4: Job transformation, not displacement, will define the next phase of automation and AI adoption.

Only 6% of U.S. employment (9.2 million jobs) faces near-term automation displacement by being both highly automated and lacking in nontechnical barriers. "Displacement is going to be a story ... but it will certainly not be the whole story, and maybe not even the majority of the story,” Ladner said.

The research shows that:

• More than 90% of jobs have some level of protection from near-term automation displacement risk.
• 7.8% of U.S. employment (12 million jobs) is at least 50% done using GenAI.

The findings underscore that AI and automation’s biggest impact on employment will come not from job loss, but from how work itself evolves.

For HR leaders, the implication is clear: Success depends on helping workers adapt to changes in technology, tasks, and the skills they possess.