AI’s Future-of-Work reality check: surprises & new strategies for leaders.

Many companies that hoped to slash costs by swapping employees for AI are hitting hard limits. Case in point: Meta reportedly explored cutting some teams by up to 60% and installing AI “agents” to handle their work, only to halt the plan when the technology underperformed. Internal data showed that while automated systems greatly increased output (code changes up 220%), actual user-facing improvements grew only 36%, and major glitches spiked 40%. The efficiency gains simply didn’t live up to the hype, forcing leadership to pause further layoffs and reconsider where humans are indispensable.

Companies that went further down the all-AI path are now walking some of it back. Swedish fintech firm Klarna, after cutting roughly 40% of its workforce in a high-profile AI automation push, found customer satisfaction plummeting. Complex or sensitive tasks proved too much for its AI customer service bots, and even engineers and marketers had to step in to handle issues the AI couldn’t resolve. By late 2025, the CEO admitted “we went too far” and quietly began rehiring human service agents to restore quality.

These reversals are cautionary tales for any leader planning aggressive headcount cuts in the name of AI. Other “AI-first” experiments have stumbled in similar ways: Ford has rehired about 350 veteran engineers (its famed “grey beards”) after finding automated design systems couldn’t match experienced human judgment. Australia’s Commonwealth Bank reversed a move to replace dozens of call-center staff with an AI voice bot when service deteriorated and complaints soared. The pattern is clear: gambling on unproven AI to fully replace humans comes with hidden costs in quality, risk management, and customer experience. Many leading organizations are learning that the most effective use of AI is not wholesale substitution of people, but thoughtfully integrating AI with human expertise.

These developments are unfolding against a backdrop of growing employee anxiety. New data show workers are increasingly uneasy about how AI is being used in their jobs and whether leadership has their backs. Glassdoor’s midyear analysis found employee trust in senior leadership has dropped to its lowest level in nine years. In the same review data, mentions of “misalignment” shot up 95% year-over-year and talk of “distrust” climbed 18%. In other words, many workers fear their leaders are out of sync with day-to-day realities, especially when those leaders cite “AI efficiency” while making cuts or changes that employees feel are mismanaged.

Anxiety about AI in the workplace is surging. Glassdoor notes “AI anxiety” in employee posts spiked by 240% compared to last year. Much of this fear stems from uncertainty: people worry about job security, fairness, and being replaced or surveilled by algorithms. Tellingly, 76% of employees now admit they use AI tools that their bosses never officially approved. This widespread “shadow AI” means workers are experimenting under the radar, likely because they either feel mandated to boost productivity or fear penalty for using AI openly without clear guidance. Such hidden adoption poses security risks and underscores a lack of open dialogue.

Even as many employees embrace AI, they are struggling with its impact on their roles and wellbeing. A Boston Consulting Group survey earlier this year revealed a “joy paradox”: over two-thirds of workers using AI report enjoying their work more, yet 41% also feel increased mental strain. And while 42% of front-line employees say they save at least one day a week thanks to AI tools, 66% receive little to no direction on how to reinvest that time productively. The result is often frustration and burnout instead of the expected boost in morale and innovation. Leaders pushing AI adoption must recognize the human side of this transformation – without trust, support, and clarity, the workforce may resist or fail to realize AI’s full benefits.

As AI capabilities race ahead, organizations face a pressing question: do we truly have an AI skills gap, or a strategy gap? Many companies cite a dearth of AI talent, but new research suggests the bigger issue is that leadership hasn’t defined a clear plan. In one 2026 survey of over 100 senior leaders, 47% said the single biggest barrier to closing their AI skill gaps wasn’t budget or employee resistance – it was the lack of a well-defined AI strategy. Over half also admitted they “can’t find” the right talent to craft that strategy in the first place. In short, many organizations don’t fully know what they want to achieve with AI, making it nearly impossible to train or hire effectively for it.

This strategic void is contributing to a growing capability divide in the labor market. Companies that have a vision for integrating AI are actively reshaping roles and upskilling their people, while those without clear direction risk falling behind. The latest McKinsey global survey found nearly 90% of large firms worldwide use AI in at least one function, and 80% of those report individual productivity gains from the tech. Yet simply deploying tools is not a guarantee of success. The variance in outcomes is huge: employees who use AI for only one or two tasks see positive productivity gains less than half the time, whereas those using it broadly for many purposes report success rates around 90%. The difference isn’t that one company bought a better bot than another; it often comes down to how well leadership redesigns workflows and trains people to truly leverage these tools.

Forward-thinking organizations are treating AI as augmentation, not automation. They are creating new hybrid roles like “AI integration specialists” and “AI ethicists,” and encouraging the rise of what Upwork dubs the “AI orchestrator” – employees who pair deep domain expertise with AI fluency to drive outcomes. Demand for these skills is exploding. According to a global analysis by PwC, jobs requiring AI skills are growing almost eight times faster than the overall job market (up 69% vs 9%). And workers with strong AI capabilities are commanding wage premiums above 50%, reflecting how scarce and valuable these skillsets have become. Organizations that empower their people with training and redesign roles around human-AI collaboration are not only finding it easier to fill critical positions, they’re also seeing productivity leap ahead. In fact, the most AI-advanced companies (the so-called “superstars”) have achieved 163% higher labor productivity since 2018, vastly outperforming peers. The big takeaway: building internal AI capability—through clear strategy, talent development and new roles—is emerging as a decisive competitive advantage.

All these trends point to one thing: the human side of AI transformation is now the choke point—and enlightened leadership is urgently needed. Leaders who get it are starting to act. At IBM, for example, the Chief HR Officer recently bucked the industry trend of slashing junior roles by announcing plans to triple entry-level hiring across all business units. Her reasoning was simple: if the company stopped hiring and training new talent now, its future talent pipeline and ability to leverage AI would dry up within a few years. IBM’s move highlights a growing recognition that long-term success with AI will require nurturing human expertise (even as entry-level tasks evolve) rather than cutting off the supply of new skills.

Executives are also learning that how they themselves use AI can send a powerful signal to the workforce. A new report this week found that when managers rely on AI to do their own jobs, 42% of employees perceive those managers as less competent, and over a third trust their judgment less. Employees want leaders to embrace innovation, but not to the point of abdicating personal expertise or insight. The implication is clear: leaders must model responsible AI use—using the tools for support and efficiency while continuing to provide human judgment, context, and accountability. If employees see their bosses outsourcing core leadership tasks to algorithms without transparency or explanation, it undermines confidence and morale when they need guidance the most.

Finally, with formal regulation lagging behind technology, stakeholders outside the C-suite are stepping in to set guardrails. Unions have scored landmark agreements to protect workers from unchecked automation and data misuse. In 2026 alone, labor groups from port operators to Hollywood writers negotiated contract clauses banning job-killing tech and mandating worker consent or training for workplace AI deployments. Such provisions, now present in dozens of union contracts, are becoming an unexpected blueprint for HR policy. Even in companies without unions, boards and executives are paying attention. The clear message is that proactively addressing employee concerns—through transparency, upskilling commitments, and ethical AI guidelines—is not just a nice-to-have, but increasingly essential for retaining trust and staying ahead of legal risks.