Artificial intelligence does not reduce jobs. According to new research, it strengthens it.
An eight-month study of how generative AI changed work habits at a US-based technology company with about 200 employees found that they worked at a faster pace, took on a wider range of tasks, and worked longer hours in the day, often without being asked.
And this was not the fault of the slave-driving employers. In fact, the company did not even require the use of AI, although it offered subscriptions to off-the-shelf AI tools.
“Workers are voluntarily doing more because AI makes ‘doing more’ seem possible, accessible, and often inherently rewarding,” Aruna Ranganathan, an associate professor at the University of California Haas School of Business, and doctoral student Shin-Chi McGee Yeh wrote in the Harvard Business Review.
They identified three main forms of intensification.
- Extending the task: Taking advantage of the fact that AI can fill knowledge gaps, employees are now taking on responsibilities that previously belonged to others. Product managers and designers started writing code, researchers took on engineering tasks, and individuals took on jobs they might have previously outsourced or avoided. They initially tackled these tasks piecemeal, but eventually expanded their scope of work. “Many people experienced cognitively enhancing effects from these tools, reducing their dependence on others and providing instant feedback and corrections along the way,” the researchers noted.
- The blurred line between work and non-work: AI has made it so easy to start tasks that individuals are now slipping small amounts of work into what used to be their breaks. You can give instructions to the AI while you’re at lunch, in a meeting, or waiting for a file to load, and let it do its magic while you do other things. This prompt actually looked more like chatting than working. However, the researchers note that over time, people took fewer natural breaks and were able to work more consistently.
- More multitasking: A new rhythm has permeated the workplace, with employees continuously managing multiple active threads at once, performing their own work and monitoring AI generation. “They did this, in part, because they felt they had a ‘partner’ to handle their workload. This feeling of having a partner gave them a sense of momentum, but the reality was constant switching of attention, frequent checking of the AI’s output, and an increased number of unresolved tasks. This created cognitive load and a sense of constant juggling, even if the work felt productive,” the researchers say.
Veteran human resources executive Ashley Goodall warns of another problem: the 98 percent problem.
With most electronic tools, you expect the results to be 100% correct. If you enter the correct numbers and use the correct functions, you won’t have to guess at the calculator. If a formula in an Excel spreadsheet is valid, it can be treated as valid when calculations are performed.
This is not the case with AI. “It’s less like an inanimate tool and more like an enthusiastic intern with only a tenuous grasp of work ethics. On the one hand, someone with boundless energy and a willingness to take on any challenge, and on the other, someone who hasn’t yet figured out that lying to make people happy is not a strategy for long-term success,” he wrote on the Substack blog.
He says the point of delegating something is that you don’t have to do it yourself. However, if you have any doubt that what is returned is not perfect, you should double check. “For delegated tasks, 98 percent good is functionally the same as 60 percent good or 40 percent good. Everything needs to be checked,” he says.
This creates even more work for you as you choose what to delegate, what not to delegate, and see how your e-intern is doing.
quick hit
- Ottawa thought leader Shane Parrish advises resisting societal pressure and adapting quickly to evidence.
- Even if it’s a great idea, it won’t sell on its own, says change consultant Greg Sattell. Three forms of power must be mastered: hard power, soft power, and network power. “Hard power coerces. Soft power persuades. Network power amplifies. Real influence comes from knowing how to combine all three,” he says.
- Productivity writer Laura Vanderkam says the number one time management rule is to set a proper bedtime so you get enough sleep.
Harvey Schachter is a Kingston-based writer specializing in management issues. He is the author of the following book, along with Sheelagh Whittaker, former CEO of EDS Canada and Cancom. When Harvey Never Met Sealag: Emails on Leadership.
