On Monday 19 January, at a roundtable of the Policy Liaison Group on Workplace Wellbeing in Parliament, employers and experts from across tech and wellbeing warned that the success of artificial intelligence at work will depend less on the technology itself and more on how organisations support people through change.
Participants, including Labour MP for Dudley, Sonia Kumar MP, agreed that wellbeing, trust and culture are not optional extras, but essential foundations if AI is to deliver sustainable productivity gains rather than anxiety, disengagement and stalled adoption.
The discussion hinged on whether AI represents a net opportunity or risk to the workplace. Evidence shared at the roundtable highlighted that organisations with strong, people-centred cultures are far more likely to implement AI successfully. Where trust is low, training is inadequate or wellbeing is overlooked, AI adoption tends to increase stress and resistance rather than performance. The group stressed that productivity gains are ultimately realised through people, not systems.
Participants also cautioned against approaches that stifle innovation through excessive or poorly targeted regulation. Instead, the discussion emphasised the role of robust corporate governance principles that encourage employers to balance profit and people – collectively addressing the UK’s chronic productivity lag. Organisations need to foster cultures where employees can safely experiment with new tools, learn by doing and remove friction from everyday work. Similarly, AI is already being used to support wellbeing directly, including tools that help employees manage financial stress and access support at scale, with the added benefit of discretion. Used responsibly, AI can enhance autonomy, reduce drudgery and strengthen mental fitness, but only where trust, transparency and clear guardrails are in place.
The discussion concluded that treating wellbeing as an add-on risks undermining the very productivity gains AI is meant to deliver. Unlike previous technologies, AI systems evolve rapidly and place new cognitive and emotional demands on employees, making minimal training approaches unrealistic and counterproductive. While participants agreed that AI is currently more about augmenting human work than replacing it, they were clear that some roles will be disrupted over time, severely in some cases. Employers and policymakers must therefore plan early for reskilling and invest in leadership capability so workers are supported as jobs change. Ultimately, embedding wellbeing, trust and psychological safety from the outset was seen as essential to ensuring AI improves job quality and performance rather than accelerating burnout and disengagement.
Gethin Nadin, Chair of the Policy Liaison Group on Workplace Wellbeing, said: “The debate about AI too often swings between hype and fear. What’s missing is a focus on how technology actually interacts with human wellbeing at work. The evidence is increasingly clear that wellbeing is not an add-on to AI adoption, it is a precondition for success.
“High-performing organisations with strong, people-centred cultures are far more likely to implement AI effectively, where its integration relies on trust, engagement and support.”
Simon Greenman, Partner and Head of AI at Best Practice AI, said: “This is probably the most complex software ever introduced into the workplace, and there is no manual for it. Expecting people to master AI with minimal training leaves employees feeling overwhelmed, nervous and disengaged. In the near term, this is much more about augmentation than automation, but success depends on whether organisations create the trust, support and psychological safety people need to engage with it confidently.”
Francesca Tabor, AI Growth Hub, said: “Employers need to train their staff to use AI and give them confidence to engage with it. That means being transparent about how decisions are made and keeping humans firmly in the loop. If AI tools are introduced to support staff, they should not be used to monitor mental health or performance in ways that undermine trust. When people disengage and ‘take their brain out of the loop’, both wellbeing and outcomes suffer.”
