Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
The use of artificial intelligence has the capacity to boost productivity but the powerful technology also puts workers’ mental health at risk as it erodes traditional ways of collaborating.
ChatGPT and other generative AI tools have become a boon to some types of work. The software, which can create sophisticated text, images and computer code in a way that mimics human abilities, enables employees to complete certain tasks faster and at a higher standard. McKinsey estimates that more than 70 per cent of companies use AI in at least one area of their business, such as sales.
“Tasks like drafting contract templates, managing procurement processes, or analysing complex data — once lengthy and time-consuming — can now be completed in hours instead of days,” says Eva McLellan, managing director at Roche Pharmaceuticals.
A study, based on the daily diaries of about 100 call centre employees at a bank in South Korea, who used an AI tool to help customers, showed that the technology improves performance and knowledge acquisition.
The research, published in the Journal of Management, however also revealed information overload — blurring employees’ focus and overwhelming them with data, which in turn impaired their performance and recovery.
Overlooking the cognitive costs and considering only the technological benefits however could potentially mask the risks associated with implementing AI, in particular when it comes to mental health.
Corporate cultures that have made collaboration and psychological safety their central pillars could be threatened as staff focus on creating synergies with a machine rather than with each other. A reliance on generative AI could shrink social networks and isolate people when they most need support, affecting employees’ motivation to work together as well as friendship and trust built up in the office.
A study published in the Journal of Applied Psychology shows that the more people interact with AI to achieve work goals, the more they experience loneliness, ultimately causing insomnia and increased alcohol consumption.
However, there is a silver lining: the research shows that employees tend to compensate for interactions with AI by seeking to support their co-workers, indicating a natural push that employers should encourage.
Most companies implement generative AI by making it available to workers and letting them experiment. This however increases the feeling of isolation and ignores the opportunities in knowledge transfer by people sharing tips to use AI across roles and teams.
With the technology becoming a permanent fixture, companies need to rethink how to create opportunities for their employees to build connections while maximising productivity and quality gains.
McLellan holds a quarterly session for her teams at Roche on best practices in the use of AI to advance corporate priorities. “By sharing learnings and collaborating on innovative ways to use AI, we can strengthen both our connections and our results,” she says. “The goal is to ensure technology brings us closer together, not isolates us, as we navigate the future of work.”
Role and job design are another approach to foster the human connection. Navdeep Arora, who advises large insurance companies on AI, reports how the businesses mitigated the effects of AI automation on underwriters as the technology replaces their traditional role in accepting and pricing risks.
“They have taken proactive actions,” Arora says, adding that underwriters who manage insurance policies are able to switch from dealing with them case by case to looking at the overall profitability of a portfolio. “This helps them advise the salespeople more closely,” he adds.
Arora cites a health insurance provider that created “family teams” of staff who interact with the same insured family (the client and their dependants), a move that the company said boosted job satisfaction and improved customer experience and healthcare outcomes.
In some cases, companies monitor AI automation to ensure it is deployed where it reduces the workload. The Silicon Valley-based entrepreneur Soulaima Gourani says she limits “the use of GenAI in tasks that require creativity or emotional depth, opting for analogue tools”. She stresses that the key is not “avoiding GenAI but using it consciously, leveraging its power without letting it define us”.
Employers can mitigate concerns by limiting the activities in which AI is used, such as for synthesising data or routine tasks. Gourani stresses that Happioh, the software developed by her company of the same name, retains manual aspects because she “believes humans should stay in the driver’s seat”.
Engagement with generative AI in the workplace can boost efficiency, but it is spreading at a pace that leaves little room for companies to design stringent guardrails. Businesses need to take a step back to ensure the technology’s gains are balanced by actions to foster social connections, collaboration and to monitor for risks of burnout.
Thomas Roulet is a professor at Cambridge Judge Business School, University of Cambridge. His book, co-authored by Kiran Bhatti, ‘Wellbeing Intelligence: Building Better Mental Health at Work’, will be published in July
