# Everyone Is Told to Learn AI. Who Gets the Chance? 2026-10-06 · Somerset County, New Jersey · Reported Feature PwC’s new workforce survey suggests a divide in who gets to participate in the future of work. AI use is rising while fewer workers say they can access the learning resources they need. PwC’s new workforce survey suggests a divide in who gets to participate in the future of work. AI use is rising while fewer workers say they can access the learning resources they need. --- Imagine two employees receiving the same announcement about their company’s exciting AI future. One gets access to a tool, practical training and permission to spend part of the workday figuring out where it helps. The other gets the announcement. Both have officially been invited to adapt, although only one has been given much to adapt with. That hypothetical is a useful way into PwC’s 2026 Global Workforce Hopes and Fears Survey. Released September 29, it describes a workplace divide that deserves more attention than another round of predictions about which jobs AI will replace. Before anyone loses a job title, employees can begin to have very different answers to a more immediate question: Does this company’s future include me? PwC surveyed 49,364 workers across 48 countries and regions and 29 sectors in May and June, weighting responses by age and gender within each geography. These are global results, with a workplace question that travels readily to the United States. Sixty-four percent said they had used AI at work during the previous year. Daily generative AI use rose from 14% to 22%. Meanwhile, reported access to needed learning and development resources fell from 59% to 51%. The arithmetic is awkward. More people are encountering machinery that asks them to learn new ways of working, while fewer say they can get the resources they need to learn. A company can announce that adaptation is everybody’s responsibility. That announcement does not supply the time or instruction required to make adaptation possible. PwC calls 56% of the workforce the “engine room”: workers whose skills are less scarce and who report relatively little AI advantage. The group does much of the daily work that keeps organizations functioning. Only about two in five say they have access to the learning and development they need. In the fuller report, PwC puts that access at fewer than 40%, compared with nearly 80% among its front-runners. The label deserves care. “Less scarce” describes a position in the labor market; it does not establish that a person’s work is unimportant. An organization can depend heavily on work it believes another person could learn to do. Calling those employees the engine room recognizes their contribution, but an engine still needs maintenance. Complimenting it while neglecting it is a peculiar operating strategy. PwC’s four groups are built from workers’ perceptions of demand for their skills and the benefits they report getting from AI. They are not the results of a technical proficiency test. Alongside the engine room and front-runners are workers using AI to gain an advantage despite having less scarce skills, and workers with highly valued expertise who have relatively little AI engagement. The divide crosses familiar job categories rather than neatly replacing them. That is where the idea of a new office class system becomes useful. The distinction could increasingly be between employees who are allowed to help shape the next version of work and employees who are expected to cope with it. A title tells you where someone sits in the current organization. Access to learning may tell you considerably more about where they can go next. The survey’s differences extend beyond whether people can finish a task faster. Daily AI users reported greater confidence in job security than infrequent users, 68% versus 57%. They were also 15 percentage points more likely to trust top management, 21 points more confident in their ability to learn new skills and 12 points more likely to ask for a promotion. Those comparisons do not establish that daily AI use caused the more positive outlook. They do show that the workers using the technology most frequently experience their prospects differently. Feeling capable of learning whatever comes next changes how a person approaches the workplace. Asking for a promotion is easier to imagine when you believe your skills are growing than when every announcement makes you wonder whether you are becoming obsolete. Confidence can therefore become part of the divide itself. Consider what it means to be asked to demonstrate a new workflow while your colleague is still trying to find out which tools are approved. One person has something visible to present. The other may be putting just as much effort into keeping existing work moving, with fewer opportunities to produce the kind of improvement leadership has decided to celebrate. If employers then treat that difference as a straightforward measure of initiative, they risk rewarding an opportunity they helped distribute unevenly. Curiosity matters. So does the ability to act on it. An employee with a full workload, limited access and no clear guidance cannot turn an enthusiastic company email into a useful experiment merely by having a better attitude. The employees already benefiting are also capable of leaving. PwC identifies 14% as front-runners, combining scarce skills with strong reported AI advantage. Within that group, 29% said they were very or extremely likely to change employers over the next year. Employers face a difficult combination: some workers see expanding options outside the organization while a much larger group struggles to find a route forward inside it. That makes a training strategy built around a handful of enthusiastic volunteers look insufficient. Their demonstrations can be helpful, and their initiative deserves recognition. But someone showing what is possible does not automatically give everyone else the conditions to try it. A workplace can become very good at displaying innovation while leaving much of its staff with little room to participate. There is also the question of how much room employees have outside work. Fifty-seven percent named economic volatility as a risk to job security, compared with 44% who cited AI taking on more tasks. Only about one-third reported having money left at the end of the month. For many workers, paying for the present remains a more pressing problem than preparing for the future. Against that backdrop, expecting people to finance their own tools or fit substantial learning around an already stretched life is a consequential choice. It favors employees who have spare money, spare time or prior experience. Employers may call the result self-motivation, but they should also ask how much of it reflects conditions that made learning easier for some people than for others. PwC’s report recommends protecting time for learning and making reskilling available without increasing workloads. Those are concrete commitments a worker can recognize. Relevant instruction, access to approved tools and a manager who makes room for practice offer more than a general encouragement to become AI fluent. They make it possible to connect learning to the job someone actually has. The question for employers is whether they are willing to treat learning as work. If it matters enough to appear in performance expectations, it matters enough to occupy time on the calendar. A future built around new skills requires more than announcing which skills will be valuable and waiting to see who manages to acquire them. AI may change what employees can do. The choices companies make now will help determine who gets to discover that possibility. When everybody is told to adapt, the clearest sign of commitment is who receives a real chance to learn, and whether that chance reaches the people keeping the place running. Sources • PwC — Global Workforce Hopes and Fears Survey 2026 and methodology • PwC release via PR Newswire — September 29, 2026 --- ProbleMattic is written and maintained by Matthew Kulcsar, a software engineer, project manager, technologist, platform builder, emergency-services-trained helper, grandfather, and lifelong collector of broken systems, odd behaviors, and useful nonsense.