Is there a right amount of AI use at work to boost happiness?

The hype surrounding AI has died down a little bit in recent times, with generative AI officially in Gartner's "trough of disillusionment". Deployment seems to be following a familiar pattern, with initial forays ad hoc efforts driven by employees, before organizations try to formalize things (often unsuccessfully).

As with most previous technologies, it's likely to be a little while before expectations settle down and processes are re-engineered to make better use of what's available.

Involved in the change

As with so much of working life, giving people a voice and a degree of autonomy around the tools they use (and the impact of these tools on their working life) is nearly always best, but this doesn't always happen, with various stories of employees being forced to adopt AI. Most notably, tech giant Meta was recently uncovered demanding that employees be assessed based on their use of AI in driving impact.

This sense that AI is no longer optional pervades the tech industry, but research from the University of Münster suggests we should perhaps think twice. The researchers wanted to understand how, or indeed whether, AI adoption affects job satisfaction at work.

"We argue that with increasing levels of adoption, the relationship between employees’ perceived benefits and costs of AI changes, resulting in an inverted U-shaped relationship between AI adoption and job satisfaction," the researchers explain.

Measuring adoption

In total, they analyzed over 500 public firms from across the United States, with earnings call transcripts used to determine AI adoption. They then used data from both Glassdoor and LinkedIn to determine job satisfaction.

Researchers looked at over 2.4 million Glassdoor reviews, counting any mentions of AI, particularly around exploration or exploitation of the technology. They also analyzed LinkedIn data to look for data governance-related jobs at each of the firms.

The results show that job satisfaction is at its highest when there is a low to moderate level of AI adoption. When adoption is high, job satisfaction actually declines.

Finding the sweetspot

The findings demonstrate the competing forces at play when we roll out AI in the workplace. There are clearly some initial benefits when we're able to automate some structured and repetitive tasks. This low-hanging fruit frees us up to do more engaging and challenging tasks. Workers gain independence, learning opportunities, and relief from drudgery.

Beneath these early gains, however, is a form of impairment effect that gathers pace as adoption levels increase. As the technology moves from truly augmenting our work to replacing it, the jobs that remain become sufficiently cognitively demanding to become uncomfortable. We face information overload, a dependence on technology we may not fully understand, and a sense that our own relevance is on the line.

"Your job becomes more complex because you have to do more to justify your position," explained one technology manager.

There's a mathematics underlying this tension. Benefits increase linearly; costs compound exponentially. The result is a satisfaction curve that rises, peaks, then falls, typically somewhere in the middle range of AI adoption.

The exploration advantage

Not all firms navigate this curve equally well. Those with an exploration orientation, such as those who emphasize experimentation, risk-taking, and continuous learning, can tolerate significantly more AI before satisfaction declines. Their employees, already comfortable with demanding tasks and constant adaptation, view AI as a liberating tool rather than a threatening replacement. The turning point shifts rightward by roughly 30%.

Interestingly, data governance appears to flow in the opposite direction. When companies have a robust data governance infrastructure, the satisfaction curve is flattened. Employees at such firms are already familiar with data-driven work, so they gain less from additional AI. What's more, they also have a better understanding of the technology, and its implications, so they feel the risks much earlier than their less savvy peers.

When discussing these two elements with managers, complexity and autonomy emerged as key job characteristics. For instance, complexity is key because workers initially loved AI's ability to take on routine work and free them up for more complex tasks. As this adoption intensified, however, it often meant that the complexity went beyond their current skill set.

Autonomy presented a similarly nuanced picture. When AI adoption was nascent, it allowed people to be more independent and not rely as much on colleagues. As AI adoption deepened, however, it took on more of the decision-making, and employees felt constrained, even neutered. Hard-won experience was suddenly secondary to whatever AI recommended.

Overly optimistic

The findings suggest the augmentation thesis that is so popular among many (ie, that AI will complement rather than replace us) is overly simplistic and optimistic. The idea of collaborating with AI is undoubtedly appealing, but the reality is that employees view the technology as both friend and foe, with the balance between the two shifting as adoption deepens.

This matters beyond worker satisfaction. Companies that push AI adoption too aggressively risk undermining the very workforce they need to compete. Job satisfaction correlates with performance, retention, and innovation. When firms put too much emphasis on AI adoption and fail to take into account the morale of the workforce, any success may be pyrrhic in nature.

The implications for managers are clear. While the temptation is to jump aboard the hype train and invest in more and more AI, the reality is that for an engaged and happy workforce, the best approach is likely to require staying below the technological frontier. It's also important to take into account the culture of the firm and the departments within it. If you're more exploration-oriented, then you can invest more aggressively in AI, but those with a strong data-driven culture should proceed more cautiously.

As artificial intelligence continues its march through the economy, companies face a delicate balancing act. The technology offers genuine productivity gains, but extracting them requires sensitivity to how work transforms under AI's influence. The firms that thrive will be those that recognize AI adoption as a human challenge, not merely a technical one.

Unlock more exclusive Cybernews content on YouTube.