Artificial Intelligence (AI) is rapidly reshaping industries, bringing efficiency and innovation to areas previously dominated by manual processes. In the construction and health and safety sectors, AI offers the promise of streamlining risk assessments, improving compliance, and transforming training.
Yet, the question remains—can AI truly replace human expertise in safety-critical tasks? Should risk assessments be generated by AI without human oversight? What are the potential risks of relying on AI-driven data for health and safety decisions?
As AI tools become increasingly embedded in workplace operations, it is crucial for organisations to understand both the opportunities and the potential dangers. This editorial explores how AI is transforming health and safety, the risks associated with poor data, and the importance of adopting secure, closed AI systems to ensure safety remains paramount.
The Role of AI in Health & Safety
AI-driven technologies such as ChatGPT, Copilot, and Gemini are increasingly being used to assist in documentation, policy creation, and process automation. These tools have the ability to analyse vast amounts of data, identify patterns, and produce structured reports in a fraction of the time it would take a human.
When applied to risk assessments, AI can help identify hazards, propose control measures, and structure documentation efficiently. The potential benefits are clear:
- AI can automate the risk assessment writing process, reducing the time spent on paperwork and compliance administration.
- AI can highlight potential hazards and suggest control measures, supporting safety professionals in making informed decisions.
- AI can be integrated into training programs, allowing learners to interact with AI-powered assistants to deepen their understanding of safety procedures.
However, while these capabilities are impressive, they come with significant risks.
The Dangers of Relying on Open AI in Risk Management
Open AI models, such as publicly available AI chatbots, are trained on vast amounts of internet data. While this allows them to generate responses on a wide range of topics, it also introduces a major issue—data quality.
- Open AI models pull information from the internet, which includes misinformation, outdated regulations, and low-quality sources.
- AI-generated content is only as reliable as the data it is trained on, meaning inaccurate or misleading risk assessments could be produced.
- AI models can experience “hallucinations,” a term used to describe instances where AI generates plausible-sounding but factually incorrect information.
For industries where safety is non-negotiable, relying on AI-generated risk assessments without expert review presents a serious hazard. The risk is not only that AI might produce flawed assessments but also that inexperienced professionals might use these assessments without realising their shortcomings.
Closed AI – A Safer Approach to AI in Health & Safety
Unlike open AI models, closed AI is purpose-built for specific applications. It is not trained on publicly available internet data but instead developed using trusted, industry-specific sources.
At www.isocomply.ai, a secure, UK-hosted closed AI system has been created specifically for ISO 45001 Occupational Health & Safety Management Systems. This AI model is trained exclusively on:
- ISO standards and regulations.
- Health and Safety Executive (HSE) guidance.
- Approved Codes of Practice (ACOPs).
- Industry best practices and case studies.
By integrating AI in this controlled and structured way, organisations can:
- Accelerate the risk assessment process while maintaining compliance.
- Ensure assessments are aligned with up-to-date industry regulations.
- Reduce the administrative burden on safety professionals, allowing them to focus on proactive site engagement.
However, while AI can significantly enhance efficiency, it must never replace human oversight. AI-generated risk assessments should always be reviewed by a competent person before being implemented in workplace procedures.
AI in Training – A New Era for Health & Safety Education
Beyond risk assessment automation, AI is now playing a crucial role in professional training. AI-powered technical assistants are being integrated into safety courses, transforming how professionals learn and apply safety principles.
This technology is currently being used to support learning in:
- IOSH Managing Safely.
- NEBOSH safety courses.
- Other industry-specific health and safety training programs.
AI-powered assistants allow learners to:
- Ask technical safety-related questions and receive detailed responses.
- Request real-world examples to understand risk assessments in practice.
- Learn at their own pace while receiving guidance from AI-driven insights.
By integrating AI into training environments, professionals gain access to a virtual mentor—enhancing their ability to understand complex safety regulations and apply them effectively in real-world situations.
AI as an Enhancement, Not a Replacement
A common concern among health and safety professionals is that AI might replace human roles in risk management. However, the reality is quite the opposite.
- AI will not replace safety managers; rather, it will act as an advanced technical assistant.
- AI can automate time-consuming documentation, allowing professionals to spend more time engaging with the workforce, contractors, and supply chains.
- AI enhances decision-making by providing rapid insights, but human expertise remains essential for evaluating and implementing safety measures.
Instead of fearing AI, professionals should view it as an opportunity to enhance workplace safety, streamline compliance, and foster a stronger safety culture.
Preparing for the Future – A Call for Industry Debate
The integration of AI into health and safety management is inevitable. The question is not whether AI should be used, but how it should be governed, structured, and implemented responsibly.
Organisations must ask themselves:
- How should AI be integrated into risk management without compromising safety?
- What policies should be in place to ensure AI is used responsibly?
- How do we balance automation with the need for human oversight?
- Are we prepared to embrace AI-driven safety training as the next step in professional development?
AI is here to stay, and its influence in health and safety will only grow. But the key to success lies in how we use it—ensuring that AI enhances, rather than undermines, workplace safety.
What are your thoughts? Should AI be used in risk assessment? How can we ensure AI-generated safety documentation is both accurate and compliant?
Join the conversation and help shape the future of AI in health and safety.
