AI’s workforce divide widens, sparking talent churn and new guardrails.
A two-speed workforce emerges.
New research from PwC’s Global Workforce Hopes and Fears Survey 2026 shines a light on how unevenly the benefits of workplace AI are being felt ([1]) ([2]). Nearly 50,000 workers across 48 countries were surveyed, and while almost two-thirds have now used AI on the job in the past year ([3]), access to training and support is not keeping pace ([4]). Daily use of generative AI has jumped by more than 50% - from 14% of workers a year ago to 22% in 2026 ([5]) - yet the share of employees who say they can get the learning and development they need has actually fallen to just 51%, down from 59% last year ([6]).
This mismatch is creating what PwC calls a “two-speed” workforce ([7]). The survey identified four distinct groups of employees in the AI era ([8]) ([9]). At one end are the AI “front-runners” (14% of workers) - those with scarce, in-demand skills who are using artificial intelligence extensively or daily ([10]). Close behind them are “AI insurgents” (18%) - staff whose skills are less specialised but who eagerly leverage AI to boost their output ([11]). Another 11% are deemed “indispensables” - employees with critical expertise who have not yet embraced AI much ([12]). Finally, the largest cohort - dubbed the “engine room” at 56% of workers - has more routine skills, limited access to AI and fewer opportunities to innovate ([13]). This majority risks being left behind as workplace technology and skill needs evolve.
Trust, confidence - and talent at risk.
One counterintuitive finding is that those using AI regularly tend to feel more secure and positive about their careers. Daily AI users are far more confident in their job security (68% of them, versus 57% of infrequent users) ([1]). They also report feeling “energised and empowered” by using AI rather than threatened by it - by a ratio of about 5 to 1 ([2]). These power-users are more likely to trust senior management and to seek promotions, suggesting that effective AI adoption can boost engagement and ambition ([3]).
Yet the same elite group may pose a new retention challenge. PwC’s data shows that nearly one-third of these "front-runner" employees are already planning to change employers in the next year ([4]) - in large part because rapid inflation and cost-of-living pressures are making them dissatisfied with current pay. In other words, companies that nurture AI talent and see strong results could still lose those very people if they don’t feel adequately rewarded. This looming flight of top talent adds urgency to closing the skills gap across the broader workforce. Without wider access to AI tools and training, more workers fear their careers could stall. And some are not passively waiting to be disrupted: surveys earlier this year found almost 30% of employees - especially younger staff - quietly resisting or even sabotaging their employers’ AI initiatives due to fears of being sidelined by technology ([5]) ([6]).
From adoption to transformation.
The mixed outcomes highlighted in recent reports illustrate a core truth: adopting AI tools is just the first step, not a silver bullet. Many organisations have heavily invested in AI, inspired by efficiency gains, yet only a fraction report meaningful productivity improvements so far ([1]). Experts warn that simply adding AI into existing structures and expecting immediate returns is a recipe for disappointment ([2]). Without rethinking workflows, roles and incentives, potential gains can be left on the table.
Forward-thinking leaders are therefore focusing on organisational transformation alongside technology deployment. The PwC study urges leaders to “rewire work” - redesigning jobs and processes so that human employees can truly leverage AI to amplify their impact ([3]). This means not just automating tasks, but also elevating human skills and redefining team practices. Companies that empower their people to collaborate effectively with AI - rather than simply replacing or sidelining workers - are seeing better returns. In fact, the survey found daily AI users often emerge as star performers, highlighting the value of training more employees to use these tools confidently.
Guardrails and governance catch up.
As businesses scramble to unlock AI’s benefits without alienating their workforce, policymakers and labour groups are stepping in to set new boundaries. On 30 September, California became the first US state to enact laws protecting workers from certain AI-driven harms ([1]). The new legislation prohibits employers from relying solely on algorithms when making hiring, disciplinary or firing decisions, requiring a real person to review any AI-prompted terminations ([2]). Companies must also notify employees in writing if an "AI system" is a reason behind mass layoffs, workplace relocations or firings ([3]). And in a move to curb invasive monitoring, the state banned AI-powered surveillance in workplaces’ bathrooms ([4]).
Labour advocates hailed these as “first-in-the-nation” safeguards and called on other jurisdictions to follow suit ([5]). Unions argue that new rules like these are necessary to ensure AI “expands opportunity - not at the expense of workers” ([6]). Meanwhile, regulators beyond California are also scrutinising how AI is implemented in workplaces. This week, for example, the US Federal Trade Commission launched an investigation into major AI firms over potential harms to consumers and labour markets ([7]). The message to employers is clear: as AI transforms work, businesses will be expected to uphold transparency, fairness and human oversight - or face growing legal and reputational risks.