How to Overcome Productivity Paranoia in the Age of AI

When, as a result of the COVID-19 pandemic, many employees began working from home full-time, numerous managers experienced what Satya Nadella, CEO of Microsoft, called “productivity paranoia.” These managers were distrustful of their employees’ productivity, who had suddenly disappeared from their line of sight. This led to a surge in the sales of so-called tattleware or bossware—software designed to monitor workers’ online activity—while in other companies, some managers tried to cope with the anxiety of the situation by overloading their employees’ schedules with an endless series of low-value meetings and video calls.

However, in contrast to their supervisors’ concerns, many employees claimed that they were just as productive, if not more so, working remotely than in the office. In fact, most said they would like to continue working remotely, even if only a few days a week, once the pandemic was over.

What employees were less comfortable with was the intrusion of electronic monitoring systems imposed by their companies. And they had good reason. We now know, thanks in large part to the 2022 meta-analysis by Siegel, König, and Lazar, that these practices not only increase stress and reduce job satisfaction but can also trigger counterproductive behaviors, especially when employees perceive that these measures prioritize the organization’s interests over their own. This explains the success at the time of devices like “mouse jigglers” or the website overemployed.com.

All those meetings and video calls weren’t well received either. For many, it made no sense. Their managers didn’t trust them to be productive when working from home, but it was precisely these managers who were hindering productivity the most with such micromanagement practices.

Fortunately, once the pandemic subsided, many companies ended up normalizing remote work, to the point where today, the number of people working remotely, at least partially, is nearly double what it was in 2019. However, this was not the case everywhere. Some business leaders, victims of that productivity paranoia, decided to implement return-to-office policies, the effectiveness of which has since been questioned by science. As evidenced by the 2023 study by Ding and Ma, not only was there no evidence that return-to-office mandates benefited business performance or shareholder value, but the study also highlighted how such measures deteriorate job satisfaction, work-life balance, perceptions of senior management, and corporate culture.

Now, with the arrival of artificial intelligence in the workplace, I fear we may witness a resurgence of this productivity paranoia, especially concerning so-called “knowledge workers.”

The first thing we need to understand is that “productivity paranoia” regarding this class of professionals does not arise because it’s difficult to measure their productivity when we cannot see how many hours they work each day (their input). The real problem lies in the fact that traditional methods of measuring productivity, which many companies still use, and which originate from the industrial era, are not suitable for evaluating the work of this category of professionals, regardless of where or when they do their work.

To begin with, as Peter Drucker said, the productivity of knowledge professionals cannot be assessed solely based on the number of “things” (output) they produce in a given period (input). Knowledge work is much more complex and open-ended than the routine tasks of the industrial era, from which the management systems many companies still use originate. This creates confusion among both business leaders and many Human Resources professionals.

Moreover, unlike routine work, knowledge work is not pre-programmed. True knowledge workers do not perform a fixed list of tasks each day. Therefore, the key question in this type of work is: “What is the task?” And that is a question that often only the knowledge worker themselves can answer.

These workers also need autonomy to explore, experiment, and learn. Continuous experimentation and innovation are essential parts of their responsibilities. Knowledge work also involves a continuous cycle of learning and teaching. Unlike repetitive jobs, it requires constant learning by the knowledge worker, as well as an ongoing teaching process, in which they share their knowledge with others inside and outside the organization.

Another important difference is that knowledge workers are not focused on maximizing the quantity produced, but on achieving quality. Yet, it’s not about meeting a minimum quality standard, as in routine jobs; it’s about achieving optimal quality. Only then can they consider the volume of work, and the effort required.

One last aspect to keep in mind when managing the productivity of knowledge workers, especially in a context where many companies claim to have difficulties attracting the talent they need, is that knowledge workers own their means of production: their ideas and knowledge, as well as their “relational capital,” and this gives them greater mobility. As Drucker pointed out, often jobs need these people more than the people need their jobs. And companies need to learn to navigate this reality.

However, despite all these differences, many bosses continue to apply management practices designed for a very different type of worker to the knowledge professionals on their teams. They struggle to accept that getting one of these professionals to come up with an idea or make a decision that positively impacts the company’s competitiveness has little to do with the number of hours the person spends in front of their computer. And this is a problem for them, their employees, and their companies.

The thing is this problem could worsen with the arrival of artificial intelligence in the workplace.

Several experiments with different generative AI solutions have demonstrated that by using these tools, people can accomplish the same work in much less time. This is partly because tools like ChatGPT have eliminated the “blank page,” significantly easing the creative process. Additionally, there is evidence that AI improves the average quality of a team’s work, particularly benefiting those who, without access to these tools, did not achieve as high a level of performance. Therefore, if companies expect the same level of output from their employees as before, the arrival of AI will allow these individuals to complete their work in less time, with less effort, and in some cases, with better quality deliverables.

These changes brought about by AI may create a feeling among managers similar to what many experienced when remote work became widespread during the COVID-19 pandemic, and their employees suddenly disappeared from their view. However, the concerns that arise now are different. It is no longer just about the fear that employees might deceive them when they are not being watched. The new challenge is that knowledge workers now have a powerful “secret weapon”—AI—that makes it even harder to assess their level of productivity.

How much less time will my employees need to complete their work with these new tools? How much will they impact the quality of their work? How can I distinguish between the employee’s merit and that of the machine? These are questions that thousands of managers around the world are likely already asking themselves.

Some companies have conducted studies to try to quantify the increase in their workers’ productivity due to the use of artificial intelligence. This can be relatively easy when it comes to routine tasks, but as we’ve seen, it’s not so straightforward when it involves knowledge professionals.

If you ask them, few will admit they have extra time, especially if they believe this could threaten their job security or lead their bosses to increase their workload, risking that work-life balance that people increasingly value. For this very reason, if a tool now allows them to generate a report in a day that previously took a week, I don’t think many people will rush to hand it over to their boss as soon as it’s ready. It’s more likely that they’ll take things more slowly. If they work remotely, it’s as simple as using a spoon to keep themselves in “Active” status on Teams or other collaboration networks their company uses, and meanwhile, they can dedicate their time to other matters of personal interest. Even if they work on-site, it won’t be difficult for them to occupy the time freed up by AI by exploring content seemingly related to their work, calling or attending meetings that could be avoided, or adding decorative details to their work that add little value to the recipients.

The risk of these dynamics is that tasks that previously added value to their organizations could end up becoming one of those “bullshit jobs” that David Graeber discussed in his book of the same title. This, again, is not good for anyone.

In this scenario, how can we prevent the arrival of AI from triggering a “productivity paranoia” among managers, even more intense than the one caused by forced remote work during the pandemic, and from leading workers to take “productivity theater” to a new level in response to their bosses’ growing concerns?

One possible theoretical solution would be to ban the use of AI in our organizations. However, this would not only mean giving up the productivity gains that these tools can offer but could also lead employees to feel restricted. If we don’t provide them with the productivity tools available in the market, which they know they could use at other companies, they might try to bring them from home, with all the cybersecurity risks that entails, or worse, they might feel that we are wasting their time, something they are increasingly less willing to tolerate.

Fortunately, there are more reasonable alternatives to prevent this situation.

First, it’s essential to understand that the productivity of knowledge professionals can only be evaluated after a tangible result has been delivered. It is at that moment when we can assess the quality of the idea, decision, or content produced and compare it with the resources employed in its creation. However, in some cases, even the result may not allow for an adequate judgment of its quality, as this may depend on the long-term effects of the decision or idea generated. On the other hand, if predicting how long it might take to generate an idea is already complicated for establishing a standard of comparison, the incorporation of increasingly powerful AI tools will make it even more difficult. For these reasons, it would be more appropriate to shift the focus from input to output and concentrate the evaluation on the quality of the results these professionals produce. To achieve this, instead of relying on traditional productivity metrics, companies need to develop new indicators that reflect the value added by their knowledge workers with the help of artificial intelligence, such as innovation, improvement in the quality of outcomes, and impact on the organization’s strategic objectives.

Trust is another key aspect of successfully implementing technologies like AI in organizations. Leaders must earn the trust of their employees by demonstrating, through concrete actions, that they will not use AI to unfairly reduce (or increase) their workload. Similarly, it’s crucial for leaders to show that they trust their employees by avoiding excessive use of monitoring tools and instead fostering an environment that values autonomy and individual responsibility. This will help prevent the “productivity theater,” with which some workers respond to their bosses’ paranoia, from obscuring the benefits that AI can bring to the organization. Additionally, a psychologically safe environment where people feel secure in sharing the productivity gains they achieve by using AI tools in their jobs can be equally valuable.

Regarding work organization, AI-driven tools may lead companies to raise their expectations, demanding that their employees produce more or reach higher quality levels than what machines can offer. However, as always, it’s important to avoid a “one-size-fits-all” approach here. It’s crucial to identify which activities can be sufficiently handled by machines and which require human intervention to add differential value. This also represents an excellent opportunity to reflect on the true value of human work.

It’s also important to highlight the value of skills and the time dedicated to learning and exploring new areas, for which a skills-based management model can be very helpful. “Flow to work” organizational models, such as project-based work, where people move to where the work is rather than the other way around, allow for better utilization of talent. Internal talent markets, which facilitate the dynamic allocation of capabilities according to the organization’s changing needs, point in the same direction, and more and more companies are daring to experiment with them. Similarly, systems that incentivize continuous learning, certification of acquired skills, and link these achievements to employability can help employees use the time they gain from AI to develop professionally, rather than filling it with activities of no value to the company.

In addition, it’s crucial to give people reasons to collaborate and contribute to collective success. AI can be a powerful tool for improving collaboration and knowledge sharing within the organization. Fostering an environment where employees share their experiences and best practices with AI can lead to collective improvements in efficiency and work quality, reducing anxiety about individual productivity. However, for this to be effective, people need a clear purpose that motivates their collaboration.

It’s equally important to prioritize employee well-being. The adoption of AI should not create unrealistic expectations about employee availability and workload but promoting healthy and flexible work practices that can help maintain a sustainable and productive work environment amid technological change. In a context where stress and other mental health disorders are on the rise, and where work occupies a different place in people’s lives, it’s vital to recognize the importance of work-life balance. In this regard, leveraging AI’s productivity gains to evolve toward more flexible work models or even reducing workers’ hours are two possible paths to explore.

Finally, it’s crucial to continuously evaluate and adjust. This new technology is advancing rapidly, and we need to understand its implementation in our organizations as a dynamic process. Regularly assessing how AI tools are being used and their impact on productivity and work quality, along with periodic adjustments and openness to continuous feedback, will help us optimize their use and address any concerns that may arise.

In short, we must accept that AI is here to stay, and we need to adapt to this new reality without falling into the paranoia we saw during the pandemic. If we do it right, not only will we improve productivity, but we will also build a healthier, more flexible, and ultimately more human work environment.

References

Drucker, P. (1959). Landmarks of Tomorrow: a report on the new postmodern world. Harper.

Drucker, P. (1999). Management challenges for the 21st century. Harper Business.

Graeber, D. (2019). Bullshit jobs: The rise of pointless work, and what we can do about it. Penguin.

Ma, M. S., & Ding, Y. (2023). Return-to-Office Mandates. Yuye, Return-to-Office Mandates (December 25, 2023).

Siegel, R., König, C. J., & Lazar, V. (2022). The impact of electronic monitoring on employees’ job satisfaction, stress, performance, and counterproductive work behavior: A meta-analysis. Computers in Human Behavior Reports, 8, 100227.