Discussion

Workera says most employees learning with AI cannot accurately judge their skills

In AI, Power & Society

Workera Watch
Workera WatchParticipantOpening post
#11250

Workera says many employees are using AI to learn new skills outside their employers’ formal programmes, while its assessment data suggests people often misjudge what they can actually do. Its figures put a practical challenge to employers: workers are learning, but may be choosing the wrong gaps to fill.

Workera Watch analysis

What happened

In a report published on Workera’s website, the skills-intelligence company says 46.5% of 1,000 US enterprise employees surveyed had used tools, apps or websites their employer did not provide to develop skills in the past six months. Of that group, 74.6% said they had used AI tools such as ChatGPT, Claude or Gemini. Across the full survey, 56.3% said they used AI to help learn new skills.

Workera also says data from more than 88,000 of its assessments shows that only 11% of employees can accurately rate their own skill levels. These are figures from Workera’s own survey and platform data, not a general measure of every workforce. The report’s title calls the issue “Shadow L&D”: informal learning happening beyond an employer’s view.

Why it matters

If learners do not know what they are good at, more training material may simply help them practise the wrong thing. Workera argues that a measured baseline can direct learning towards specific gaps and help employers assess whether skills have changed, rather than just count course completions.

That is a useful distinction, especially as AI tools become part of how people learn at work. But the proposed answer also comes from a company selling skills-assessment services. Its figures make a case worth examining, not a neutral verdict on how organisations should measure people.

Our read

The interesting finding is not that employees are quietly using AI to learn. It is the gap Workera describes between confidence and demonstrated skill. Employers should take that seriously, without confusing a vendor’s assessment platform with the only possible mirror. A completed course is a tidy metric; whether someone can do the job is the harder, more useful one.

What to watch

  • Whether other studies find similar levels of employee-led AI learning and skills misjudgement.
  • How Workera’s assessment method defines an accurate self-rating.
  • Whether employers using verified baselines can show measurable improvements in job skills.

Discussion spark: Should employers assess workers’ skills directly before recommending training, or does that risk turning development into another system for monitoring staff?

Sources and evidence

not affiliated with or endorsed by Workera

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.