Can AI Support LOLER and PUWER Inspections?

Artificial intelligence or AI is finding its way into many areas of our day-to-day work life. In the field of equipment safety there is growing interest in how these technologies might support the inspections required under the Lifting Operations and Lifting Equipment Regulations (LOLER) and the Provision and Use of Work Equipment Regulations (PUWER). Supporters argue that AI can capture data more efficiently, flag potential issues earlier and reduce human error. Critics warn that over-reliance could create blind spots and legal risks. The reality is likely to sit somewhere between these two positions. AI is not a replacement for a competent person but it may prove to be a valuable tool when deployed carefully and with proper governance. 

The Regulatory Baseline 

The Health and Safety Executive (HSE) makes clear that statutory duties remain unchanged. This means that Thorough examinations under LOLER must continue to always be performed by a competent person. The examination report must also include the information set out in Schedule 1 such as the date, the defects found and the next due date, and records must be kept in a secure form that cannot be altered without authorisation – legal responsibilities which cannot be delegated to software or an algorithm. 

Warnings from Industry Voices

Over the past couple of years, industry bodies have also raised concerns about over-reliance on generative AI. The Consolidated Fork Truck Services (CFTS) and other sector voices have warned that machine generated outputs can be misleading, incomplete or overly general. The Project Safety Journal has also reported similar concerns.  

In a world where AI is being used to support mundane takes, it is important to highlight that businesses must not mistake AI generated text for legal compliance advice or use it as a substitute for the expertise of a competent person. 

Where AI Can Add Value 

Despite the need for caution, AI is beginning to show real benefits within the industry. Several organisations are now using drones and computer vision systems to gather images of equipment and detect signs of wear or corrosion. National Grid’s VICAP project is one example, where autonomous drones flown ‘beyond visual line of sight’ (BVLOS) are used to gather high-definition close-up images of pylons, which are then processed using artificial intelligence (AI) to show the health of the steelwork. 

Other promising applications include predictive maintenance using sensor data, automatic logging of inspection images with secure time stamps, and digital audit trails that strengthen record keeping. All of these can reduce the burden on inspectors and improve consistency, provided the competent person remains in control of the final judgement. 

Essential Governance 

For AI tools to be used safely they must be validated to demonstrate accuracy and acceptable false negative rates. The raw data and outputs must be stored securely with full audit trails and staff must be trained to interpret results and understand when human intervention is required. Liability and insurance cover must also be reviewed so that businesses are not exposed if the AI fails to detect a defect. Regulators and insurers are already asking for transparency, explainability and clear evidence that AI is being used only as a support tool and not as a statutory replacement. 

Looking Ahead 

AI will never remove the need for competent persons under LOLER and PUWER, but it can become a valuable partner in capturing data, prioritising risks and forecasting deterioration. Organisations that explore these tools now, while keeping human inspectors at the centre, will benefit from efficiency gains without stepping outside the legal framework.