RiskCompli

← All insights

Industry News By SafeCompli · 21 August 2026

HSE on AI at work: treat it like any other safety-critical kit

Originally published on SafeCompli, our parent brand. Reproduced here for RiskCompli readers.

HSE’s message on AI at work is almost disappointingly sensible.

If AI affects how people work safely, smart glasses barking instructions, cameras spotting missing PPE, software directing vehicles, treat it like any other safety-critical kit.

Not like a sci-fi sidekick. Not like a free pass. Like a bit of equipment that can get it wrong on a busy Tuesday.

What HSE says

On 10 June 2026, HSE published its regulatory approach to artificial intelligence.

Health and safety law is largely goal-setting under the Health and Safety at Work etc. Act 1974. It applies regardless of the technology in use, including AI in the workplace.

Where AI impacts workplace health and safety, you still need a risk assessment and appropriate controls so far as is reasonably practicable. That should include addressing cyber security threats.

HSE’s aim is to reach a point where AI risk is no longer novel and is managed in the same way as any other risk.

What that looks like in practice

If the system can change how people move, decide, or get told what to do, ask the unglamorous questions:

Before you buy the clever camera that “solves PPE compliance”, ask the boring question first: if this thing fails, misreads, or gets gamed, who gets hurt and how do we stop that?

Source

Have you got any AI or “smart” kit on site yet that actually changes how people move, decide, or get told what to do, or is it still mostly PowerPoint?

If you’d rather get HSE notes in your inbox than wait for the algorithm to invent a new reason to hide them, the list is here: https://safecompli.co.uk/subscribe?src=linkedin

On a side note: you can now try RiskCompli FREE, no card needed: riskcompli.safecompli.co.uk

www.safecompli.co.uk