TECH NEWS – In this case, AI didn’t take people’s jobs (it did for 7,800 people at IBM, as we wrote yesterday). It helped them become better.

ChatGPT, Midjourney, Dall-E, and many other AIs capable of creating content are making significant inroads, putting artists, for example, in a scared position (they fear losing their jobs, and that’s why screenwriters are currently on strike in Hollywood). MIT and Stanford University researchers published a study in which an unnamed Fortune 500 company implemented a unique ChatGPT model to help people who specialize in IT assistance. The study was run in late 2020 and early 2021. The results were surprising: not only did employees’ performance improve, but also their retention, as they were less likely to leave (although the turnover rate, the measure of how much the workforce changes from year to year, is relatively high at 60%!)

In the process, customer service agents talked to users in one chat window while ChatGPT was open in another. When the user asked for a solution to a software issue, the agent asked the chatbot this question and applied the result to solve the problem. “At a technical level, customer support is well-suited for current generative AI tools. From an AI’s perspective, customer-agent conversations can be considered a series of pattern-matching problems in which one looks for an optimal sequence of actions. [There is] suggestive evidence that AI recommendations can make low-skilled workers communicate more like high-skilled workers,” the study wrote.

The chatbot was trained on the company’s highly rated telephone helpline, and thanks to it, it was able to list solutions for employees and use specific phrases while doing so, so that a less IT-savvy worker could bring it up to the level of a more professional customer service agent: recruits were able to handle more calls per hour, and customer satisfaction was also much higher, with fewer instances of the phrase “I’d like to speak to your manager,” which usually hurts the employee…

Veteran and better-trained employees found the chatbot less valuable because they already knew the questions’ answers. Erik Brynjolfsson, an economist at Stanford and a researcher on the study, said people needed two months of experience to get to the six-month level. The rub is that workers with artificial intelligence can get less money ready because of a lack of knowledge…

Source: PCGamer