AI has rapidly become part of everyday working life from drafting documents to analysing data and supporting decision‑making. Many of us use it daily (I even used it to help shape this article).

I am no AI expert, but I delved a bit deeper due to becoming increasingly aware of AI generated reports, chatbots supporting mental health and AI advising workplaces on adjustments. I am not totally against the use of AI where it can support healthier, safer and speedier outcomes for people, but, I do worry about the risks to employees and organisations of not using AI safely.

As a vocational rehabilitation consultant working closely with my team of occupational health (OH) clinicians, I’ve been exploring the pros and cons of AI on workplace health. Here are the key insights and what you can do next.

AI is creating new pressures – often quietly

AI is reshaping roles, expectations, and workloads. Employees are being asked to use AI tools without training, clarity, or boundaries. This is creating several risks aligned with the HSE Management Standards, and each one requires a proactive HR response.

1. Role ambiguity and the rise of “AI shame”

Many employees are unsure when to use AI, how much to rely on it, or what “good” looks like. A growing number describe AI shame, which is feeling embarrassed about:

  • using AI “too much,”
  • not using it well enough, or
  • worrying it exposes a lack of skill or intelligence.

This quietly erodes confidence and psychological safety.

What HR can do:

  • Define clear expectations for when and how AI should be used
  • Update job descriptions to reflect new responsibilities
  • Communicate boundaries around decision‑making and oversight
  • Create open dialogue and regular feedback to build trust and confidence

2. Increased demands – AI isn’t always reducing workload

Research shows that AI can intensify work (especially in medicine and health where it is not currently improving productivity due to the way in which it is being used). Instead, it often adds steps including supervising outputs, correcting errors, learning new systems and managing new workflows.

What HR can do:

  • Review job design to ensure AI isn’t increasing workload
  • Monitor cognitive load and screen‑time fatigue – DSE assessments may need updating to reflect the increased use of AI and what this means for someone’s job tasks and workstation use
  • Assess workload regularly as AI tools are introduced

3. Poorly managed change – AI is moving faster than people can adapt

AI is often rolled out quickly, without consultation or training. This can create anxiety, resistance, and stress.

What HR can do:

  • Provide structured training rather than assuming digital confidence
  • Involve employees early in decisions about AI use
  • Communicate the “why” behind AI adoption

4. Psychosocial strain – fear, uncertainty, and loss of control

Employees may worry about being replaced, judged by AI outputs, or left behind. These concerns affect wellbeing, engagement, and performance.

What HR can do:

  • Include AI within stress risk assessments
  • Create psychologically safe spaces to discuss concerns
  • Reassure staff about human oversight
  • Support managers who are also adapting to new tools

Can AI replace workplace health assessments?

In short: no.

AI can help generate ideas for workplace adjustments or support plans but it cannot replace specialist Occupational Health (OH), Vocational Rehabilitation (VR) or ergonomic assessments.

There are real risks if AI is used to:

  • suggest reasonable adjustments
  • recommend mental health support or management of long term health conditions
  • draft return‑to‑work plans
  • interpret clinical information

I find it really frustrating when I read an external report which I am 99% certain is AI generated…it feels totally impersonal and could be potentially dangerous. AI may miss nuance, misunderstand context, or suggest actions that are inappropriate, unsafe, or legally risky. And crucially, employees lose the space to feel heard and explore what will genuinely help them.  I believe you only get that in a confidential, safe space with another human.

What HR can do:

  • Always involve a specialist (OH, VR, ergonomics, mental health) before agreeing any adjustment
  • Use AI only as a brainstorming tool, not a decision‑maker
  • Ensure clinical or professional oversight for anything health‑related
  • Document human review of any AI‑generated suggestions

This protects employees and protects the organisation.

Where AI can add real value in workplace health

When implemented thoughtfully, AI can support workplace health by:

  • Reducing admin time and workload pressure
  • Analysing wellbeing or absence trends to put support in place early
  • Supporting transcription, translation, and summarising
  • Making the workplace more inclusive through being an assistive tool for people with health conditions or neurodivergences
  • Drafting notes or reports (with clinical oversight)
  • Providing early triage and signposting

The key is intentional, ethical, human‑centred use.

Final thoughts

AI has enormous potential but only if we implement it in ways that protect wellbeing, support good job design, and maintain human connection. Leaders and HR teams have a crucial role to play in shaping this future.

Further reading: