- ChatGPT is fine for ideas, a first list of hazards or rewording your notes
- It can misquote or invent regulations, so every reference needs checking
- It has no sign-off and no record of who was briefed
- swiftRMS drafts a formatted RAMS to UK legislation, ready to review and sign
- Either way, a competent person must check it before it's used
Side by side
swiftRMS vs ChatGPT for RAMS
A general chat assistant compared with a tool built to write and sign off UK RAMS.
Starts from
- swiftRMS
- Your description of the job and the site
- ChatGPT
- A blank chat
UK legislation
- swiftRMS
- Cited for the task, such as CDM 2015, HASAWA 1974, COSHH 2002 and PUWER 1998
- ChatGPT
- Only if you ask, and it can misquote
Risk scores
- swiftRMS
- Likelihood and severity, with a 5x5 matrix on the PDF
- ChatGPT
- Whatever format it picks that day
Nearest A&E
- swiftRMS
- Found for the site, with a route map and QR
- ChatGPT
- You look it up yourself
Your company's rules
- swiftRMS
- Site rules, PPE and emergency arrangements from your Standards (Business)
- ChatGPT
- Pasted into every chat
Finished document
- swiftRMS
- Branded PDF and Word
- ChatGPT
- Text you copy into a template
Sign-off
- swiftRMS
- Manager sign-off, workers sign by QR on Business
- ChatGPT
- ChatGPT: no
Record of who was briefed
- swiftRMS
- Sign-off register as CSV or PDF (Business)
- ChatGPT
- ChatGPT: no
Cost
- swiftRMS
- From £49 a month
- ChatGPT
- Free, or a ChatGPT subscription
ChatGPT's answers vary with the version and the prompt. This compares a general chat assistant with a RAMS tool, not one particular prompt.
Fair to ChatGPT
Where ChatGPT Helps
Brainstorming hazards
Ask what can go wrong on an unusual task and it gives you a useful starting list to check against your own experience.
Rewording your notes
It's good at turning rough notes into clear sentences for a method statement step or a toolbox talk.
Explaining a regulation
It can summarise what a regulation covers, as long as you check the summary against legislation.gov.uk or hse.gov.uk.
Simple, low-risk jobs
For a routine task, with an experienced person checking every line, a careful prompt can save time.
Where it goes wrong
Five Risks of Using ChatGPT for RAMS
Wrong or made-up references
Chat assistants can cite a regulation that doesn't exist or quote the wrong year, such as CDM 2007 instead of CDM 2015. A principal contractor will spot it.
Generic hazards
It doesn't know the site, the access, the public nearby or the plant you're using unless you type all of it in, every time.
No company rules
Your site rules, PPE standard and emergency arrangements have to be pasted into every chat, and it may still leave them out.
No sign-off or briefing record
A chat transcript isn't a signed document. There's no approval, no record of who read it and no signatures from the crew.
Your name is on it
If the RAMS is wrong, it's your document and your liability, not the chatbot's. Whoever signs it must have checked it properly.
How swiftRMS does it
From Job Description to Signed RAMS
Describe the job
Type what you're doing and where, in plain words. You can add site photos too.
Get a full draft
swiftRMS writes the hazards, risk scores, controls, method statement, PPE and the UK legislation that applies, with the nearest A&E and a route map. A draft takes about 2 minutes.
Review and edit
A competent person in your business reads it, changes what's wrong and adds what only you know about the site.
Sign it off
A manager signs, confirming they've read it and reviewed the AI content themselves. On Business, a high-risk document also needs a named co-signer.
Brief the crew
On Business, workers scan the QR code on the document, read it on their phone and sign. No login needed.
FAQ
Frequently Asked Questions
Can ChatGPT write a RAMS that a principal contractor will accept?
Sometimes, if an experienced person writes a detailed prompt and checks every line. The usual reasons a ChatGPT RAMS gets sent back are generic hazards, wrong or out-of-date regulations, and missing site details such as access, welfare and the nearest A&E. A RAMS tool builds those in, but a competent person still has to review the document.
Is it legal to use AI to write a risk assessment?
Yes. The law cares that the risk assessment is suitable and sufficient and that a competent person was involved, not which tool drafted it. Under the Management of Health and Safety at Work Regulations 1999 the employer is responsible for the assessment, so whoever signs it off must have checked it properly.
What's the difference between ChatGPT and an AI RAMS generator?
ChatGPT is a general assistant that answers whatever you type. An AI RAMS generator such as swiftRMS is built for one job: it asks for the details a RAMS needs, writes to a fixed structure with UK legislation and risk scores, produces a branded PDF and handles sign-off. The structure, checks and record around the AI are what you're paying for.
Does swiftRMS use ChatGPT?
No. swiftRMS uses Google's Gemini models, set up to write UK RAMS with British English and UK legislation. Our privacy policy explains what is sent to it.
Is a ChatGPT RAMS good enough for a small job?
For a simple, low-risk task, a RAMS drafted with ChatGPT and properly checked by someone competent can be fine. The risk grows with the job: work at height, hot works, confined spaces and anything near the public need site-specific detail and a clear sign-off, which a chat reply doesn't give you.
Try It on Your Next Job
Describe the job you'd have typed into ChatGPT and compare what comes back. 14-day free trial, no card needed.