While AI is taking on work across the economy, supervising AI agents is likely to have hidden challenges and these demands must be quantified, acknowledged and built into roles, say researchers from Microsoft and Imperial College London.
Published in the Society of Occupational Medicine’s (SOM) journal Occupational Medicine, researchers say AI is likely to make accessing workplace health support much easier for employees and managers. One example of this is with AI automating and simplifying booking processes and appointments.
They say clinicians will increasingly use AI tools to manage, track and analyse workplace health and workplace health data in their organisations.
However, as this technology becomes ever more deeply embedded within workplaces, it will simultaneously create its own new health issues and challenges.
Until now, much of the conversation around AI has been about the extent to which it will replace people’s jobs. Researchers writing in Occupational Medicine believe that, for many, it will become a technology that simply exists alongside their role.
They say this approach will change how many jobs work and the likely demands on workers. While the ‘drudge’ of many day-to-day roles will be taken away by AI, human roles will become more focused on handling and managing complexity.
“As AI absorbs routine tasks, human roles may shift toward stewardship, problem-solving or emotional labour, all with their own psychological demands,” the research team, led by Imperial clinical research fellow Dr. Lara Shemtob, said.
“Individuals who currently operate independently may be expected to manage a number of AI agents and move through iterations of management practices as organisations and the technology they deploy evolve with time.
“It is essential that the demand of supervising AI is quantified, acknowledged and built into roles as they evolve to avoid hidden workloads that negate the benefit of automating outsourcing tasks.”
There will also be challenges around AI ‘hallucination’ – where AI generates results that are inaccurate or misleading. The research team say this may escalate and become harder for human supervisors to detect as the technology becomes more autonomous, putting extra stress and burdens on those humans working alongside AI tools.
As the sophistication of AI accelerates and the tasks it can be given – and taken away from human co-workers – increases, workplaces will see growing role ambiguity, potentially creating uncertainty, stress and anxiety.
“AI will alter the roles, risks and responsibilities of the human workforce across all the sectors that occupational health supports,” the team adds. “Understanding and managing the interface between humans and AI is therefore the next critical frontier for occupational health, and one upon which we must bring our expertise to bear.”
