American workers just hit a 39-year job satisfaction peak — and we at The Conference Board have the data to prove it. At nearly 69%, it is the highest level of job satisfaction we have recorded since we began tracking in 1987.
It may also be the high-water mark. Because the same force most likely to define the next decade of work — AI — is already threatening to pull that number back down for a large portion of the workforce.
AI is improving work dramatically for some employees while leaving others feeling less secure, less confident, and further behind. If organizations aren’t careful, their approach to AI will split their workforce into haves and have-nots, undermining the technology’s benefits.
Our research found that employees using advanced AI tools are significantly more likely to report higher job satisfaction, engagement, mental health, and intent to stay. Nearly 40% of workers say AI has improved their job satisfaction, and the share saying it has significantly improved satisfaction rose sharply over the past year. But that leaves the majority of the workforce somewhere between indifferent and actively anxious — and anxious is where the story gets alarming.
AI has had a mixed impact on career expectations. While 30% of employed workers say AI has made them more confident in their career prospects, nearly a quarter say it has made them less confident. Among unemployed adults looking for work, the anxiety is even more pronounced: They are almost twice as likely to say AI hurts their prospects. Add the neutral and the worried together, and you have a workforce in which a large portion of workers are not benefiting from the AI moment that is supposedly transforming everything.
This should be a warning sign for business leaders. Workers who feel optimistic about AI’s effect on their careers report dramatically stronger workplace outcomes across engagement, effort, and intent to stay. Workers who feel threatened by AI score more than 25 percentage points lower on these measures. That 25-point gap is not a rounding error. It is the distance between a workforce that sustains a historic satisfaction peak and one that squanders it.
Organizations may be unintentionally creating a two-tier workforce: employees whose confidence, productivity, and opportunities accelerate with AI, and employees who increasingly feel left behind. The record high we documented is an average. Averages hide fault lines.
The warning signs become even clearer when looking at who benefits most from AI. Men are more likely than women to report positive effects from AI on both their job satisfaction and
career confidence. Higher-income workers are far more likely than lower-income workers to report positive AI experiences. Workers with access to training, managerial support, and experimentation opportunities are more likely to thrive than those left to navigate AI on their own. The unlucky half is not random. It skews female, lower-income, and undertrained — the same workers who already had the least cushion.
The companies that benefit most from AI will be the ones that ensure all workers feel empowered by it. Access to tools, training, and guidance cannot be treated as perks. They need to be part of the basic infrastructure of work.
That starts with pairing AI access with meaningful skill building and support. Giving people access to technology without guidance is more lottery than strategy. Workers who thrive with AI tend to be those whose organizations actively encourage experimentation and invest in building skills, not just buying licenses.
But this divide cannot be managed if it’s not measured. Many organizations track overall AI adoption, but few examine how the experience differs across gender, income levels, or functions. To mitigate divisions, leaders should look more closely at whether AI adoption is improving work for some employees while eroding it for others. They should also be asking a sharper question: are we protecting the conditions that produced a 39-year satisfaction high, or are we quietly dismantling them?
Leaders must also recognize that AI adoption is just as much an employee experience issue as it is a technology challenge. Employees who believe that AI strengthens their careers report higher engagement, loyalty, and discretionary effort than those who believe AI weakens their prospects. When workers believe AI makes them replaceable, it may drive disengagement, lower morale, and eventually attrition.
As a broader backlash against AI gains steam across America, corporate leaders must gain workers’ trust if they hope to implement AI successfully. This means using it to improve the work experience wherever possible, including through augmentation rather than replacement.
But the issue goes beyond culture and job security. A workforce divided in its ability to benefit from AI also hurts business performance. Teams evolve at different speeds, innovation concentrates among a small subset of employees, and trust across the organization weakens.
We spent nearly four decades tracking the slow, hard-won rise of American job satisfaction. AI can extend those gains — or it can concentrate them so narrowly that the aggregate number stops meaning anything. The organizations that succeed will be the ones that treat that peak as something worth protecting, not just a baseline to optimize around.
Because once a workforce splits into AI haves and have-nots, rebuilding organizational trust becomes far harder than deploying technology in the first place.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
