As of 13 August 2026, AI can only partly create health and safety training for UK employees.
This still needs a person who signs their name to it.
Can you do it?
15 minutesto a draft.
1 hourto something you’d act on.
Cost, all in£0
Skill neededpower-user
Who has to check ita colleague
What the alternative costsColossyan is an AI avatar video tool aimed at workplace training content, although no price is provided here.
If this goes wrong, employees may be taught procedures that do not match the workplace and an incident may follow.
What to actually do
Hand it to a person
The route this page recommends
A person who owns the outcome does this end to end, worth it when the failure is dear.
Use a tool built for this
Second choiceDo it yourself
The distant thirdA chat interface, power-user skill, and roughly 1 hour until you can act on the result.
How to actually do it
- Open your current health and safety policy, relevant risk assessments, safe systems of work, emergency procedures and equipment instructions, and remove personal data that the training does not need.
- Ask the person responsible for workplace health and safety for the target roles, required completion date, delivery format, previous incidents or near misses that may be used, and the name of the competent person who will check the material.
- Paste the documents and workplace details into a chatbot with the copyable prompt, then ask it to produce the complete training package and list every missing fact.
- Compare each procedure, hazard and control in the draft with the current workplace documents, correcting any difference against the approved version rather than accepting the model's wording.
- Ask the competent health and safety person to check the legal references, risk controls, emergency instructions, role-specific scenarios and any required practical demonstration.
- Put the approved content into your chosen format, deliver it to the employees, record attendance and completion, and keep the final approved version with its review date.
Prompt
Create a UK workplace health and safety training package for [organisation and job role]. Use only the workplace information and documents I provide, and clearly label anything that is missing or needs confirmation. Do not invent hazards, controls, legal duties, incident figures, emergency arrangements or equipment details. Produce: 1) learning objectives, 2) a logical session plan with timings, 3) plain-English teaching notes, 4) a slide-by-slide outline, 5) a short knowledge check with answers, 6) practical scenarios relevant to the role, 7) a trainer script, 8) an attendance and completion record, and 9) a list of points that must be checked by our competent health and safety person before use. Keep the content suitable for UK employees and distinguish general guidance from our organisation's own procedures. Do not diagnose injuries or illnesses. If the material refers to urgent or worrying symptoms, direct people to NHS 111 or emergency services as appropriate rather than giving a diagnosis. Source any legal or official claims from current UK official guidance and identify the source for each one. Workplace information and documents: [paste here].
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
What it gets wrong
- AI cannot inspect the workplace or know whether the written procedure matches what employees actually encounter.
- AI cannot decide whether a risk assessment or control measure is adequate for a particular task.
- AI cannot replace a competent person's check of legal duties, emergency arrangements or practical demonstrations.
- AI cannot confirm that employees understood the training or can perform a safety-critical task correctly.
- The employer still carries responsibility for unsuitable training, even when AI drafted it.
What caps this at PARTLY: legal accountability, judgement under ambiguity and stakes of error.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 1 |
| Verification | 1 |
| Liability | 1 |
| Effort delta | 2 |
| Total | 7 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can ChatGPT create health and safety training?
- Yes, it can draft lesson plans, slides, trainer notes, quizzes and video scripts. It cannot know whether the content matches your premises, equipment or risk controls, so a competent health and safety person must check it before delivery.
- Is AI health and safety training legally compliant in the UK?
- AI cannot certify that training is legally compliant. Use current UK official guidance and your own risk assessments, then have the relevant competent person confirm that the training meets your duties and workplace arrangements.
- Can AI make a health and safety training video?
- Yes. Tools such as Colossyan can help turn an approved script into workplace training video content, but an avatar video does not prove that employees understood the procedure or can carry it out safely.
- Can AI give health and safety advice about an injury?
- Do not use a chatbot to diagnose an injury or decide whether urgent treatment is needed. For urgent or worrying symptoms, use NHS 111, and call emergency services when the situation is life-threatening.
Nearby answers
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.