Original LinkedIn edition published
These are original companion notes. Read the full newsletter on LinkedIn
The short answer
Self-directed AI learning is useful, but the route someone took to learn does not establish what they understand or how they use that knowledge at work. Use a common knowledge baseline when needed, then look at task practice and organisational support separately.
Measure learning without mistaking the route for the result
Employees often build AI capability through experimentation, colleagues and tutorials before a formal programme catches up. That initiative is an asset. It is also difficult to assess through licences, prompt counts or a training register.
A common Core-6 knowledge baseline can show whether people understand AI foundations, risks, data boundaries, verification and oversight. It cannot establish transfer to a particular workflow. L&D teams can separately give people safe opportunities to practise with feedback, then observe whether the learning holds in new cases.
This distinction helps avoid two errors: discounting self-taught people because they lack a certificate, and assuming that a certificate or confident account proves applied capability. Start with the decision the evidence is meant to inform, then choose methods suited to it.
Illustrative example: an L&D cohort
Several employees have taught themselves to summarise long internal reports with AI. Rather than counting prompts or licences, the L&D team checks common knowledge on data boundaries and output evaluation. It then separately lets participants practise on approved material, with feedback on whether they verify claims and protect confidential information.
Questions to take into your next meeting
- Ask what the learning needs to enable in a real workflow.
- Measure foundational knowledge without treating usage frequency as a score.
- Offer safe task practice with feedback, separate from the Core-6 assessment.
- Check whether tools, guidance and managers support the behaviour expected.
Put the idea to work
Try this with a learning cohort
Before the next AI course, choose one knowledge area to baseline and one real work behaviour to observe separately. Review both afterwards. A change in knowledge is useful evidence, even when behaviour needs more time and support to change.
What would show that informal learning has become dependable practice in your organisation?
Sources and further context
- European Commission AI literacy Q&A Context-specific literacy measures and internal records of support
Core-6 measures common AI knowledge through Sova. The activities discussed here are separate ways to explore work and behaviour.
Assessment versus confidence survey Try the separate A/A/H and PRIME worksheet See what Core-6 measures