It is no secret that the world is in the midst of substantial AI transformation. The health and safety (H&S) space is no exception, with AI tools becoming increasingly available and progressively more useful in assisting to achieve compliance with H&S duties.

Indeed responses received by the Health and Safety Executive (HSE), in preparing their Understanding how AI is used in HSE Regulated Sectors Report (2025) (HSE Report), indicated that AI is being used in a span of industries including agriculture, commercial consumer services, manufacturing, logistics and transport, public services and construction.

Whilst these tools may provide an ever more cost and time efficient approach, and improve workplace safety, their adoption should not be devoid of consideration for ethical concerns and legal obligations which may arise. This article explores both sides this coin.

How can AI assist with H&S matters?

AI tools can be used for a spectrum of H&S related tasks, ranging from compliance to hazard and risk identification to H&S training.

1.    Automated monitoring of H&S processes and reporting

AI tools can provide a way to monitor H&S processes and enable automatic alerts of gaps in information and process, upcoming deadlines and/or outdated information. This in turn can assist duty holders to stay up to date with internal H&S reporting and compliance requirements.

The ability of AI to monitor and review data automatically can reduce the chance of human error, which can arise when inputting data or keeping track of deadlines. For example, AI may be used to monitor the schedule of plant maintenance and generate automatic alerts when a particular piece of plant is due for routine maintenance, without the need to rely on an onsite employee or the H&S team to manually track this schedule. This is especially useful for companies that operate over multiple sites and have a vast amount of plant that may be on differing maintenance schedules.

Utilising AI in this way can also reduce the administrative workload of H&S teams and may enable them to dedicate more time to addressing complex H&S issues and workforce engagement.

2.    Identification of hazards and risks

Various AI tools can assist in identifying and monitoring patterns of events or incidents which in turn can assist with identifying hazards and/or anticipating future areas of risk. Further, they can be utilised to complete risk assessments.
For example, an AI tool that monitors the number of hours employees in a particular team are logging during the workweek, can assist in alerting relevant personal to the risk of employees having excessive workloads and working excessive hours – both of which can contribute to an increased likelihood of psychosocial hazards arising. 
Another example are tools which monitor, in real time, an individual’s facial and eye movements in order to identify signs of fatigue. Such tools are, at times, used in the long-haul transport industry in order to mitigate against hazards that can arise when a driver becomes fatigued. In circumstances where signs of fatigue are identified, the system alerts drivers to indicate that a break should be taken. These tools can also provide data to employers to assist them in identifying areas of risk with certain routes or if there is a need for H&S training or communications to alert employees to related hazards and risks.

3.    Enhancing H&S training

AI can also be utilised to make H&S training more engaging or workplace specific, which in turn may improve participants’ attention and enjoyment during training sessions, and their understanding of the H&S content. As an example, tools that produce realistic simulations can be utilised as a way to generate workplace specific scenarios, which employees are then required to work through whilst not being exposed to the actual hazards or risks.  Alternatively, AI can also be utilised to generate scenarios that form part of H&S training material with workplace H&S data able to be input in order for the scenarios to be applicable to that workplace.

As AI becomes more accessible, from a cost and resource perspective, it may be that its use in providing H&S training may be increasingly regarded as reasonably practicable – a key consideration for employers in respect of their general H&S duties to their employees.

4.    The communication of H&S information

Finally, the usefulness of AI in communicating complex information, in engaging and accessible ways, is becoming increasingly evident. The use of such AI tools may be especially helpful in assisting employers to ensure that the necessary H&S information is provided to employees – as they are required to do so under their general duty to employees under the Health and Safety at Work etc. Act 1974.

Firstly, as most people would be aware, there are multiple AI tools that can prepare written text following the input of key data and requirements such as tone. These tools can be especially helpful for H&S teams that may have limited time to spend on preparing workforce communications or if there is a large and diverse workforce for which communication style is best adapted depending on the specific audience.

Secondly, AI tools that involve automated chat functions can provide a way for employees to quickly and easily access relevant information and ask questions to enhance their understanding of H&S matters, as they arise. Again, such tools can be especially useful for employers with large workforces for which a variety of H&S matter may arise at varying times. Of course, as with the use of any AI tool, the use of AI in this way would need to be properly established using accurate information and consistently monitored to ensure that any H&S information provided to employees is appropriate.

What to keep in mind when using AI tools

Whilst AI may have cost, time and compliance benefits it would be unwise, and in some cases unlawful, to employ workplace AI tools without first considering and addressing ethical risks, and other legal obligations that can arise.

In relation to the use of AI in the H&S space specifically, these matters which require further consideration include:

•    Privacy: Users of AI in the workplace should ensure that these tools are used in a way that does not infringe on employees’ rights including rights to privacy. For example, if utilising AI to monitor workplace patterns and behaviours, consideration should be had for the need to obtain that information, how it will be used and how it will be stored or destroyed.

•    Employee morale: Whilst AI can be effective in identifying H&S risks, through monitoring workplace behaviour and patterns, it is important that employees are appropriately informed of its use in order to avoid them feeling like their employer is constantly watching them perform their work. Informing employees of the AI tools being utilised may be done through inductions, workplace updates, workplace policies, and employment contracts. The risk of workers experiencing stress due to AI monitoring that adopted ‘algorithmic management’ was potential risk, from the use of AI in the H&S space, that was identified by respondents to the survey reported on in the HSE Report.

•    Use of personal information: Employers should carefully consider any personal information that is input into AI platforms, and ensure that they are handling this information in a way that is compliant with data protections laws including the United Kingdom’s General Data Protection Regulation.

•    Bias and discrimination: Due to the way in which AI tools are built and operate, there is a risk of their input, and then their output, being impacted by biases and prejudices that exist within society. This may result in discriminatory outcomes and, in the H&S space, situations that do not adequately address hazards and risks. For example, if a workplace, that is traditionally dominated by one gender, inputs historic data into an AI tool which is assisting to identify hazards and risks, the output of that AI may not be accurate when/if the gender diversity of that workplace changes.  ‘Flawed or biased data’ leading ‘to unreliable safety decisions and missed hazards’ was another risk, arising from the use of AI in the H&S space, identified by respondents to the HSE’s survey which was reported on in the HSE Report.

•    Inaccuracies: Again, due to the way that AI is developed, operates and continuously learns, if incorrect or incomplete information is input into an AI tool this will impact the quality of its output which may lead to negative consequence if that output is relied on when making H&S related decisions.

Do not ‘set and forget’

As a final point, AI tools should not be implemented with a ‘set and forget’ mentally. 
Whilst these tools may provide automated processes, they should only be used with regular human oversight to ensure that they are operating as intended, their output is accurate, and that any required updates or alterations are made to ensure that positive H&S outcomes are achieved.

A failure to appropriately implement and monitor AI tools, may itself lead to those tools giving rise to H&S hazards and risks. To mitigate against this occurring, employers should ensure that any adoption of AI is thoroughly considered, reviewed and recorded.