Can AI create my RAMS? The pros, cons and compliance considerations

Artificial intelligence is becoming an increasingly popular tool within health and safety. With the ability to generate documents in seconds, many businesses are turning to AI to save time and improve consistency. But while AI can be a useful starting point, it should never be viewed as a replacement for competent health and safety planning. 

Recently, I reviewed an AI-generated Risk Assessments and Method Statements (RAMS) document for the installation of CCTV systems. At first glance, it appeared comprehensive, covering the installation process, identifying common hazards, and including control measures.  

However, a closer review highlighted a crucial lesson: AI can help create a framework, but it cannot replace site-specific knowledge, professional judgement, or legal responsibility. 

Where AI can add value 

Used correctly, AI can be a valuable tool when developing RAMS. 

Creating a strong starting point 

One of AI’s biggest strengths is producing a clear structure. In the CCTV installation example, the document covered the full scope of work, including: 

• Site surveys 
• Bracket installation 
• Drilling and cable routing 
• Electrical connections 
• Testing and commissioning 
• Client handover 
• Waste removal 

This provides a useful foundation and ensures key stages of a task are not overlooked. 

Identifying common hazards 

AI can also help identify many of the typical hazards associated with a task. For CCTV installation, it recognised: 

• Work at height 
• Electrical hazards 
• Hidden services 
• Power tools 
• Manual handling 
• Public interface risks 
• Adverse weather 
• Lone working 

For routine activities, this can help prompt discussions that may otherwise be missed during early planning stages. 

Producing logical method statements 

A good RAMS should guide operatives through a task step by step. AI is capable of generating logical and easy-to-follow method statements that take workers from arrival on site through to completion and handover. 

Highlighting basic control measures 

Standard controls such as PPE requirements, housekeeping expectations, first aid arrangements, ladder safety guidance and emergency procedures can all be incorporated effectively by AI. 

These strengths make AI a useful assistant, particularly for generating first drafts and reducing administration time. 

Where AI falls short 

The challenge begins when AI-generated documents are treated as complete and ready for use. 

In reality, many AI-produced RAMS contain generic content that may look convincing but lacks the detail required to manage risks effectively. 

Site-specific information cannot be automated 

The most significant issue with many AI-generated RAMS is that they are not site-specific. 

A RAMS must consider: 

• The exact site location 
• Building type and construction 
• Access restrictions 
• Members of the public and other occupants 
• Emergency arrangements 
• Local rules and permit systems 
• Named supervisors and responsible persons 

AI cannot carry out a site visit, understand local conditions, or identify unique risks at a particular location. 

Without this information, a RAMS may meet a template requirement but fail to provide meaningful protection. 

Risk assessments need more than “High, Medium, Low” 

Many AI-generated assessments use simple risk ratings without explaining how those ratings were calculated. 

A robust risk assessment should include: 

• Severity scores 
• Likelihood scores 
• Residual risk calculations 
• Clear justification for risk reduction measures 

AI can suggest controls, but it cannot reliably determine actual risk levels without accurate site information and professional judgement. 

It still requires a competent assessment 

The CCTV RAMS reviewed relied heavily on ladder use and included standard control measures such as securing ladders and maintaining three points of contact. 

However, it failed to answer critical questions such as: 

• Why are ladders being used instead of a podium, tower scaffold or MEWP? 
• What is the working height? 
• How long will work be carried out at each position? 
• Is side-loading likely during drilling? 
• What rescue arrangements are in place following a fall? 

These are decisions that require competent assessment. AI can suggest controls, but it cannot justify access choices. 

Hidden risks are often missed 

Even when hazards are identified, AI often lacks the depth needed for effective risk management. 

In the CCTV example, important considerations that were largely absent included: 

• Asbestos management 
• Dropped objects 
• Water ingress following drilling 
• Fire stopping requirements 
• Network and security disruption 
• Vehicle movements 
• Welfare arrangements 
• Permit-to-work requirements 
• Noise, vibration and dust exposure 
• Fire exit obstruction risks 
• Defined weather limits 

These omissions are common because AI tends to work from patterns rather than genuine understanding of the work environment. 

The biggest risk: False confidence 

Perhaps the greatest danger is that AI-generated RAMS often look impressive. 

Well-written content, professional formatting and technical language can create the impression that risks have been fully considered when important gaps remain. 

A RAMS should never be judged by how polished it looks. It should be judged by whether it accurately reflects the work, the site and the controls needed to keep people safe. 

The best approach: AI + human expertise 

We believe AI has an important role to play in health and safety. It can speed up document creation, improve consistency and remove unnecessary administration. 

However, every RAMS should still be reviewed by a competent person who can apply site knowledge, challenge assumptions and ensure controls are suitable for the task. 

The most effective approach is to use AI to build the framework, then apply professional expertise to make it safe, compliant and site specific. 

Make RAMS easier with SafeWorkforce 

Creating quality RAMS doesn’t have to be time-consuming. SafeWorkforce’s RAMS Builder helps businesses create professional RAMS quickly and consistently, while giving you the flexibility to add the site-specific details, controls and risk information that AI alone simply cannot provide. 

Whether you’re producing RAMS for routine maintenance, installations, construction activities or contractor management, our tools and expert support help ensure your documentation is practical, compliant and fit for purpose. 

If you’re looking to simplify RAMS creation without compromising safety, speak to the experts at SafeWorkforce and discover how our RAMS Builder can transform the way you manage risk.