Workplace risk management is increasingly defined by what can be identified early, enabling organisations to anticipate stress, fatigue and unsafe conditions before harm occurs.
Getting your Trinity Audio player ready...
null

When it comes to health and safety, many businesses continue to respond after incidents occur rather than seek to proactively anticipate them. According to the Health and Safety Executive (HSE) data from its 2024/25 report, an estimated 1.9 million UK workers are experiencing work-related ill health.1 Much of this is preventable with the right data and approach.

AI can play a key role within an organisation's Health and Safety management. Gallagher's 2026 AI Adoption and Risk Benchmarking report indicates that 63% of organisations have already implemented or operationalised AI. Increasingly, AI is reshaping how organisations identify and manage risks before they materialise, turning safety from hindsight into foresight and supporting more predictive, insight-led decision-making.

With this opportunity comes a new layer of complexity, accountability and potential risk exposure.

“
The core challenge for business leaders is recognising that while AI can enhance safety protocols, accountability remains inherently human.
Andy Northcott, commercial director, Gallagher

AI is reshaping the health and safety landscape, but its operational limitations require the same level of scrutiny as its benefits.

Where AI is transforming health and safety today

AI is now embedded in day-to-day safety practices across industries, with applications expanding rapidly in areas where real-time data can help prevent harm.

For instance, computer vision and sensor-based technologies enable earlier hazard detection on shop floors and construction sites. These systems can continuously analyse posture, monitor interactions between pedestrians and moving machinery and provide real-time feedback on manual handling. By identifying unsafe conditions and triggering anti-collision alerts, they can help reduce the risk of both musculoskeletal injuries and workplace accidents.

On the roads, AI-enabled fleet monitoring systems provide real-time analysis of driver behaviour, tracking fatigue, distraction and stress indicators throughout a journey to support safer driving practices. By aggregating and analysing this data, organisations can move beyond isolated incidents to identify broader risk trends.

These capabilities allow earlier identification of potential risks and more timely intervention. In many environments, AI is already helping teams shift from reactive incident management to more proactive risk mitigation. However, increased visibility doesn't eliminate risk. AI identifies patterns and biases but doesn't resolve them. Its value ultimately depends on how organisations interpret and act on these signals.

AI has shown merit in strengthening the foundations of health and safety frameworks. It can support:

  • Risk assessments
  • Policy development
  • Training programme design

But AI outputs are often generalised. They require interpretation and alignment with real-world conditions that algorithms cannot fully capture. Workforce diversity, operational nuance and external variables, such as weather or supply chain disruption, remain beyond its predictive certainty.

Misconceptions and the realities of accountability

One of the most persistent misconceptions is that AI is a 'magic wand' that provides definitive answers across all scenarios. This mindset creates operational vulnerabilities, including over-reliance on automated systems, the misapplication of generic data to specific local hazards and a false sense of security among management teams. In practice, AI outputs require scrutiny, validation and contextualisation by qualified professionals.

Importantly, technological advancement doesn't alter the legal landscape. Legal and operational responsibility for workplace health and safety continues to sit with business owners and directors. Regulators, including the HSE in the UK, support the adoption of automation but continue to expect robust, human-led risk assessments, regardless of the technologies in place. AI supports compliance frameworks, but it doesn't transfer liability.

Dual impact: Tension between benefits and risks

The integration of AI into health and safety frameworks presents a dual impact, simultaneously addressing long-standing challenges while introducing new risks linked to automation, data integrity, cybersecurity and the need for effective human oversight.
Where AI improves safety Where AI introduces risk
Real-time monitoring of driver behaviour, fatigue, stress and operational distress indicators. Over-reliance on automation, leading to degradation of human situational awareness.
Early risk identification through detection of complex, hidden data patterns. Silent failures or incorrect outputs, including hallucinations where AI presents inaccurate information as fact.
Consistent documentation and structured processing of compliance workflows. Cyber vulnerabilities that can compromise integrity of physical safety systems.

Workforce wellbeing and trust

Beyond physical safety, AI is also influencing workforce wellbeing.

Advanced systems can identify stress signals through behavioural patterns and communication cues. They can also provide an alternative channel for employees who may be reluctant to raise concerns directly. In some cases, this creates a more accessible space for individuals to express concerns and seek guidance.

However, the monitoring of employee metrics introduces important ethical and privacy considerations. Continuous surveillance can increase workplace stress and erode organisational trust if employees feel under constant scrutiny. Clear transparency around how AI is deployed is central to maintaining employee engagement and trust.

Ethical and sensible use of AI

Realising the benefits of AI without compromising trust depends on responsible deployment.

This is reflected in:

  • Clearly communicating how AI is used
  • Avoiding over-dependence on automated outputs
  • Robust governance and oversight mechanisms

Despite growing adoption, research indicates that only 43% of organisations have introduced a formal AI risk management framework. This highlights a critical gap between implementation and governance.

Leaders need to evaluate where AI adds value and where human judgement remains crucial. Risk assessments can be enhanced through AI-driven tools, supported by oversight from experienced practitioners. The greater risk lies not in the technology itself, but in its misuse without proper guardrails. The future of workplace safety will depend on balance, leveraging AI's capabilities while maintaining strong human oversight.

How Gallagher can help

Gallagher brings together a dedicated and experienced team of health and safety specialists within its broader risk management team. This is supported by a range of customised solutions to meet evolving organisational needs. We work with organisations to structure comprehensive risk assessments, refine corporate policies and develop training frameworks that combine technological innovation with practical, human-led risk management.

Source

1"HSE Statistics 2025: A Complete International Guide to Workplace Health and Safety Data," Arinite, 26 Mar 2026.

 

Disclaimer

The sole purpose of this article is to provide guidance on the issues covered. This article is not intended to give legal advice, and, accordingly, it should not be relied upon. It should not be regarded as a comprehensive statement of the law and/or market practice in this area. We make no claims as to the completeness or accuracy of the information contained herein or in the links which were live at the date of publication. You should not act upon (or should refrain from acting upon) information in this publication without first seeking specific legal and/or specialist advice. Arthur J. Gallagher Insurance Brokers Limited accepts no liability for any inaccuracy, omission or mistake in this publication, nor will we be responsible for any loss which may be suffered as a result of any person relying on the information contained herein.