.jpg?w=651&ssl=1)
As a result, the company spent a lot of time training new workers hired to replace those who quit. Many of the skills needed were what the researchers called “tacit knowledge,” experiential know-how that can’t be easily codified but that large language models can absorb from chat logs and then mimic. The company’s bot helped with both technical and social skills, pointing agents to relevant technical documents and suggesting chipper phrases to soothe seething customers, such as “happy to help you get this fixed asap!”
After the bot started helping out, the number of issues the team resolved per hour jumped 14 percent. What’s more, the odds that a worker would quit in a given month went down by 9 percent, and customers’ attitudes toward employees also improved. The company also saw a 25 percent decline in customers asking to speak to a manager.
But when the researchers broke the results down by skill level, they found that most of the chatbot’s benefits accrued to the least-skilled workers, who saw a 35 percent productivity bump. The highest-skilled workers saw no gain and even saw their customer satisfaction scores dip slightly, suggesting that the bot may have been a distraction.
The value of that high-skilled work, meanwhile, multiplied as the AI assistant steered lower-skilled workers to use the same techniques.
There’s reason to doubt that employers will reward that value of their own accord. Aaron Benanav, a historian at Syracuse University and author of the book Automation and the Future of Work, sees a historical parallel in Taylorism, a productivity system developed in the late 19th century by a mechanical engineer named Frederick Taylor and later adopted in Henry Ford’s car factories.
Using a stopwatch, Taylor broke physical processes down into their component parts to determine the most efficient way to complete them. He paid special attention to the most-skilled workers in a trade, Benanav says, “in order to be able to get less-skilled workers to work in the same way.” Now, instead of a fastidious engineer toting a stopwatch, machine learning tools can collect and disseminate workers’ best practices.
That didn’t work out so hot for some employees in Taylor’s era. His methods became associated with declining incomes for higher-skilled workers, because companies could pay less-skilled employees to do the same kind of work, says Benanav. Even if some high performers remained necessary, companies needed fewer of them, and competition between them increased.
“By some accounts, that played a pretty big role in sparking unionization among all these less-skilled or medium-skilled workers in the 1930s,” Benanav says. Some less-punitive schemes did emerge, however. One of Taylor’s adherents, the mechanical engineer Henry Gantt—yes, the chart guy—created a system that paid all workers a minimum wage but offered bonuses to those who also hit extra targets.
Even if employers feel incentivized to pay high performers a premium for teaching AI systems, or employees win it for themselves, dividing the spoils fairly might be tricky. For one thing, data might be pooled from several workplaces and sent to an AI company that builds a model and sells it back to individual firms.
