- Possible justification
- A learner can have a worthwhile interest in developing understanding, judgment or a new direction even if AI can already produce the output better. Protecting a learning pathway can serve that interest, provided it offers something useful and wanted. This is a reason to examine the learner’s opportunity, not to preserve every old task, qualification or human teaching method.
- Strongest objection considered
- Preserving entry tasks can turn repetitive work into a supposed educational duty, delay useful automation or protect professional gatekeepers. A capable AI tutor or simulator may provide better practice and feedback than a busy person. Some learners may prefer income, shorter hours or a different pathway. Training must not become a condition for dignity or a way to blame people for lost work.
- Who might benefit or be excluded
- Distinguish the trainee, training employer, instructor, colleagues and people served by the workplace. Ask whether pay, accessible support and practical learning opportunities reach those with fewer resources. People unable to obtain a training place, those who leave and unpaid learners may be absent from the evidence. A funded place or benefit to a provider is not automatically a benefit to the learner.
- Resources and costs to investigate
- Count purposeful practice, explanations, feedback, materials, learner effort, supervision, pay and cover separately. Productive work may also teach; labelling a task routine does not settle whether it is useful practice. Someone must absorb any displaced work. AI can change those costs and the most useful learning activities, so yesterday’s staffing pattern is not an automatic benchmark.
- An alternative to compare
- Compare a protected opportunity for varied, useful practice and feedback with output-led discretionary assignment. Keep the learner group, learning goals and available capable-AI assistance comparable; measure costs before asserting equal budgets. Separately compare workplace learning with AI-supported simulation, income support or another wanted route. No such trial has been undertaken here.
- What to check before using this elsewhere
- A proposal needs an institution responsible for the learning objective, a clear duty, usable resources, a route to question unsuitable assignments and an accountable response. Identify what the learner can actually choose. Existing German duties do not settle another jurisdiction’s law, the best teaching method or how much human instruction is justified.
- Evidence that could change the assessment
- Evidence of freely chosen, useful capability development at acceptable cost would strengthen the case. Coercion, exclusion, poor learning value or a better accessible alternative would weaken it. Follow learners who struggle or leave, and compare real options; a credential, finished task or satisfied response alone cannot prove durable understanding.
These are HBC’s judgments and questions for further research. They do not show actual benefits or costs, establish workers’ preferences or authorise us to speak for affected people.