52% of L&D Leaders Say AI Use Is Eroding Workers' Human Skills
General Assembly's survey of 510 L&D and HR leaders across six countries finds 52% say AI use degrades communication and collaboration skills — even as 83% say human skills are gaining value.

Why it matters
- 52% of 510 surveyed L&D and HR leaders in six countries say AI use is degrading skills like communication and conflict resolution.
- 83% say human skills are becoming more valued in their organisations; 59% expect soft-skills training demand to rise.
- L&D leaders fully linking strategy to business priorities were twice as likely to expect significant budget increases (23% vs 11%).
- Only 47% rank business outcomes among top success measures; 36% still count course completions.
- 57% expect growing demand for role-specific AI training; 15% will deprioritise out-of-the-box third-party courses.
More than half of Learning and Development leaders — 52% — say AI use inside their companies is actively degrading employees' human skills, including communication, conflict resolution and collaboration. That figure comes from new research published November 19 by General Assembly, which surveyed 510 HR and Learning and Development leaders at large companies across six countries.
The number lands at an awkward moment. Companies are pouring money into AI upskilling while quietly discovering that the same tools may be hollowing out the skills that are hardest to automate. The survey captures that tension from inside the training departments charged with managing it.
What did the survey actually find?
The headline finding is the skill erosion number, but it sits alongside a cluster of related pressures the survey quantified:
- 52% of respondents said AI use is degrading skills such as communication, conflict resolution and collaboration.
- 41% said one of their biggest challenges in the coming year will be ensuring employees retain or develop critical thinking, communication and other soft skills despite using AI.
- 64% said their teams are under pressure to deliver AI upskilling across the workforce.
- 83% said human skills are becoming more valued within their organisations.
- 59% expect demand for soft skills training to rise.
Those last two numbers deserve attention. Even as AI spreads, the respondents — the people running training at large firms — report that human skills are appreciating, not depreciating, in organisational value. The problem, as 41% of them see it, is that AI use threatens the very skills their organisations increasingly prize.
The survey also asked what Learning and Development teams themselves will need. AI literacy ranked most often as the top skill those teams must acquire. When asked about leading challenges, 33% cited closing the AI skills gap, while 31% pointed to change management and employee fatigue around AI.
Why does this matter beyond the training department?
The stakes run wider than course catalogues. If AI adoption weakens communication, judgement and collaboration at scale, companies face a productivity paradox: tools deployed to speed up work may degrade the interpersonal and reasoning skills that determine whether work is any good. General Assembly's data suggests large employers are aware of the risk and are not yet sure how to manage it.
There's also a workforce-policy dimension. The survey signals that training demand is bifurcating — AI skills on one track, human skills on the other — and that L&D budgets, already stretched, must cover both. With 64% of teams reporting pressure to deliver AI upskilling across entire workforces, the risk is that soft-skills training gets squeezed out by the AI mandate it's competing against.
How is training itself changing?
Training priorities are shifting from basic awareness toward targeted, role-specific learning, the research found. The expected changes are concrete:
- 57% of leaders expect demand for role-specific AI training to grow.
- 56% expect their organisations to expand personalised learning journeys.
- 55% expect more hands-on training.
- 52% expect more job-based or role-based training.
Standard external courses appear to be losing ground. Out-of-the-box third-party training was the option most likely to be deprioritised, with 15% of organisations expecting to place less emphasis on it.
Daniele Grassi, Chief Executive Officer of General Assembly, said employers are trying to balance technical training with broader workplace development. "L&D professionals are tasked with building role-specific AI skills while making sure the human skills that matter most don't erode in the process," he said.
He described the shift as a move toward integrated, practical training rather than standalone AI courses. "Increasingly, we're seeing companies looking to integrate AI and soft skills training to prepare employees for the future of work. Training is becoming more hands-on, with opportunities to practice critical thinking, judgement and collaboration alongside AI usage and get feedback in real time," Grassi said.
That model — practising judgement alongside AI usage with real-time feedback — is a direct response to the erosion problem. If AI use weakens critical thinking in daily work, the survey suggests employers want training that rebuilds it in the same setting where the erosion happens.
Who gets the budget?
The research points to a clear financial pattern. Among respondents whose learning strategy is fully linked to documented business priorities, 23% expect a significant budget increase — more than double the 11% of other respondents who expect the same.
The same advantage appeared among leaders who said they work closely with functional heads to map future skills needs, and among those who said they help shape business strategy. In other words, L&D teams embedded in business planning are winning the budget argument.
Influence at the top appears broad. 73% of respondents said company leadership regularly asks for Learning and Development input on workforce capability, transformation or future-skills decisions. 63% said L&D leaders help shape business strategy. For a function long criticised as a cost centre, those numbers suggest real strategic traction — at least among the large companies surveyed.
Can anyone prove training works?
Here the picture darkens. Even with growing influence, many L&D teams struggle to demonstrate results. Connecting learning initiatives to business outcomes or return on investment was the second most-cited challenge for the coming year, named by 35% of respondents.
The measurement gap is stark. Fewer than half — 47% — said business outcomes such as revenue growth or profit rank among their top three measures of success. Meanwhile, 36% still count course completions among their leading metrics, and 26% said distinguishing AI course completions from actual competency is a major challenge.
The divide is sharpest between specialist and generalist roles. Half of respondents in dedicated Learning and Development roles said course completions are a top metric, compared with 32% of those in broader HR roles who also oversee training — a reversal of what one might expect, with the specialists more attached to the cruder measure.
Grassi said the pressure to show results is intensifying as organisations spend more on AI-related learning. "Businesses are investing in AI upskilling, but the big challenge for L&D leaders is proving that what they're doing is working," he said.
The fix, in his view, is a change in what gets counted. "That means going beyond counting how many people took a course, and measuring behaviour changes, skills and competencies instead — and linking those to real business outcomes," he said.
What happens next?
The survey's internal logic points one direction. The organisations most likely to win bigger budgets are those tying learning to business outcomes, yet fewer than half of L&D leaders currently measure success in business terms — and a quarter can't tell an AI course completion from actual competency. As AI spending grows, the teams that close that measurement gap will be the ones funded to solve the skill-erosion problem the rest are still counting courses about.
Source: itbrief.com.au
More from Elena Vasquez
Show full bio
Market editor covering technology and product news in artificial intelligence and machine learning at AI In Context.
744 articles
Related articles
- Pearson to Acquire AI Skills Platform Workera in H2 2026
- 67.8% of Workers Now Use AI Weekly — But 56.4% Get No Time to Learn It
- Gartner: 27% of Australian workers fear AI will make their skills obsolete
- APA Tells Employers: AI Deployment Is a Job Design Problem, Not a Purchase
- 92% of HR Leaders Say Staff Can Learn AI Skills — Yet Half Hire Externally