Guest blog by Dr Lara Shemtob
Across industries, many organisations are encouraging their entire workforces to experiment with and use AI. Their agenda relates to staying competitive by improving organisational productivity and preventing a skill gap emerging.
What does the evidence tell us about workforce health and productivity with AI adoption?
Interim findings from research conducted by a team at the Haas School of Business, UC-Berkeley suggests that AI adoption can increase risk factors for work-related stress. The researchers shared early findings in a recent article for Harvard Business Review.
Research context
The team studied how generative AI changed work habits at a US based technology company with around 200 employees. The organisation facilitated access to AI tools. AI use was not mandated but the report indicates considerable voluntary adoption. The study was carried out for eight months and researchers gathered data through a combination of in-person observation, tracking communication channels and in depth interviews.
Research findings
The interim findings shared have crossover with risk factors for work related stress. The findings included:
- AI enabled workers to take on tasks not traditionally within their domain - resulting in changes of role and widening of job scopes. This also impacted colleagues’ workplace demands in peer reviewing work and mentoring individuals who were using AI to push beyond their existing skillset.
- Working with AI resulted in employees doing more work. This related to prompting AI leading task initiation to feel quicker and easier. However, objective data suggested this blurred boundaries and resulted in longer workdays overall.
- Working with AI resulted in more context switching between tasks, increasing cognitive load. Workers reported feeling they were doing more at once and experiencing more pressure when using AI than before adoption.
Why is this relevant to occupational health?
Occupational health is a clinical speciality centered on the relationship between work and health. The early findings from this research indicates working with AI can alter the relationship between work, health and productivity and increase multiple factors for work related stress (Demand, control, role, support, relationships and change).
The researchers were concerned about the sustainability of working with AI as a result of their findings. They have suggested that organisations should establish their own practices for working with AI safely and sustainably to avoid these negative impacts.
Occupational health must be part of taking an evidence based and preventative approach to building a safe and effective AI practice within organisations. The Health and Safety Executive management standards can be a useful blueprint for building practices to contain and manage risk factors for work-related stress.
Fictional Example
Scenario: Access to AI for enterprise tools are being introduced to all workers at a boutique management consultancy firm with 200 employees. The workforce is encouraged to experiment with and use AI in their work.
Workforce education: Occupational health is involved from the outset of the adoption process, educating the workforce on the risk factors for work-related stress and indicating how AI can exacerbate these.
Policy co design: The organisation designs AI usage policies with occupational health input to take account of evidence based approaches, for example relating to the HSE management standards. This might include guardrails on expected output increases, clear accountability pathways for AI augmented work and management/ supervisor training for colleagues overseeing others’ AI outputs.
Data analysis: Occupational health data is audited for sickness absence and presentations including work-related stress and burnout. This helps map change in workforce health to AI exposure patterns.
Case management: Occupational health supports individuals with ill health arising at work in relation to AI use through clinical case management. This includes providing risk assessment and assessment of fitness for work, treating clinician liaison and communication with the organisation with employee consent.
Organisational feedback loops: Where trends emerge on work-related ill health in association with AI use, occupational health supports departmental level education, intervention and audit to address factors beyond individual employees such as ‘domino effects’ of demand generated by AI use. Occupational health feeds back to senior organisational leaders on the health and productivity impact of AI implementation, and continues to work at an organisational level to iterate on existing policies, including through adapting policies to new technologies being implemented.
Takeaways: The strategic role of occupational health in workforce AI implementation
We must apply what we know about safe and healthy working practices to AI as a new technological context. To access AI’s full potential, organisations must be proactive about the work and health impact of this technology within their operations, including:
- Treating AI implementation as organisational change
- Integrating work and health expertise from the outset
- Monitoring psychological impact alongside productivity output metrics
- Adjusting policy in real time
This is aligned with a preventative agenda, moving beyond reactive sickness management to proactive work design.