Why Generative AI Might Be Making Work Harder Than Before
281Many people thought that generative AI tools, or genAI, would make work easier. Instead, some workers are feeling more drained and disconnected. They also find it harder to think deeply. This seems counterintuitive, but there’s a reason behind it. Companies rushed to adopt these tools to boost productivity, often without fully thinking about how they fit into daily routines. Now, workers are experiencing something called “prompt fatigue”—a kind of mental exhaustion caused by the constant need to craft and refine prompts for AI systems.
The Shift From Finding and Assembling to Query and Refine
Leslie Joseph, a principal analyst at Forrester, explains that work has traditionally been about “find and assemble.” You search for information, gather it, and put it together. This process is straightforward and familiar. But with large language models (LLMs), the approach has changed to “query and refine.” Now, workers ask the AI questions, get answers, and then have to tweak their prompts to get better results.
This back-and-forth can be frustrating. The AI knows a lot, but it isn’t always reliable. Workers find themselves going over the same questions repeatedly, trying to get the right answer. This breaks the natural flow of work and can make tasks feel more tedious rather than easier. Interestingly, this problem seems to be especially noticeable among software engineers. It’s unclear if other professions are feeling the same way yet.
Prompt Fatigue and Its Impact on Productivity
Ramprakash Ramamoorthy, who researches AI at ManageEngine, points out that prompt fatigue comes from three main challenges. First, choosing which AI tool to use can be confusing, especially when multiple tools are needed. Second, figuring out the right prompt to get what you want takes time. Third, workers often spend a lot of effort refining prompts without getting the perfect answer.
Most LLMs also tend to be overconfident. They don’t admit when they don’t know something. Instead, they tend to give an answer that seems confident but might be wrong. This can lead workers down the wrong path without realizing it, adding to their frustration.
While AI has the potential to help, it can sometimes slow people down. A study by Model Evaluation & Threat Research found that experienced developers actually worked 19% slower when using AI, even though they believed they were working faster. That’s a warning sign that AI isn’t always making work easier, especially for those with more experience.
There’s also a concern about how heavy reliance on AI might hurt deep thinking. Aaron McEwan from Gartner suggests that shortcutting to answers can weaken our ability to learn and develop expertise. When people rely too much on AI, they might forget how to think critically or solve problems on their own. Ramamoorthy compares this to how we used to memorize phone numbers before smartphones took over. Now, many don’t remember the numbers at all, and the same might happen with drafting responses or solving complex problems.
To avoid this, Ramamoorthy recommends that users recognize when they’re no longer thinking independently. They should clearly define when AI support is helpful and when it’s not. Using different tools for different types of questions can also help. For example, one AI for general questions and another for technical issues.
Beyond individual productivity, there are broader risks for organizations. McEwan highlights that industries like law are already facing challenges. As AI tools replace some tasks, junior lawyers may not develop the critical skills needed for advanced work. Without proper training, there’s a risk that new professionals won’t gain the experience they need to become effective senior lawyers.
McEwan emphasizes that productivity isn’t just about doing things faster. It’s also about creating value. If work is completed quickly but isn’t meaningful or accurate, it doesn’t benefit the organization. Companies need to think about how AI impacts talent development and whether employees are gaining the skills they need for future roles.
The Social Cost of Relying on AI
Another concern is how AI use affects workplace relationships. Julia Freeland Fisher, research director at the Clayton Christensen Institute, warns that over-relying on AI can weaken “weak ties”—those distant colleagues we don’t work with every day but who are still part of our network.
Fisher explains that AI offers convenience and instant answers, reducing the need for casual conversations or quick check-ins with coworkers. While this might seem efficient, it can lead to less access to important information and fewer opportunities to build social capital. Over time, this could make teams less connected and less innovative.
Cross-team collaboration often sparks new ideas. If AI encourages siloed work, organizations might lose the diversity of thought that drives innovation. Fisher warns that when we turn to bots for quick answers, we risk losing the rich interactions that help organizations grow and adapt.
In the end, while AI tools have many benefits, they also bring new challenges. Workers and companies need to be mindful of how these tools are integrated into work routines. Balancing the efficiency gains with the need for deep thinking, skill development, and social connection is key to making AI a truly helpful partner in the workplace.
What do you think?
It is nice to know your opinion. Leave a comment.