AI bots join discussions to take notes; they also point out that we interrupt and hog the discussion

From The Wall Street Journal:

Josh Stir knew he had been talking for a long time during a recent virtual company meeting. But he didn’t expect a robot to call him out on it.

Stir was presenting from his Fort Wayne, Ind., office about a new software feature that would allow his colleagues to automate tedious tasks, such as copying and pasting data. Then a notification popped up on his laptop telling him he’d been talking nonstop for 30 minutes without letting anyone else say a word.

“It was like, monologue!” says Stir, 46 years old, who works for a tax services company as a senior software development manager. “And I was like, yes, that’s what I’m here to do.”

Workers around the world are adopting artificial intelligence to streamline tasks ranging from email writing to product development. Now companies have begun using AI to root out another workplace inefficiency: meetings. Across the U.S., some workers are using tools that record, analyze and summarize what has been said, allowing them to skip gatherings entirely and skim the highlights.

The AI also acts as a kind of virtual Miss Manners, reminding people to share the mic and to modulate their speaking pace, and advising them how to avoid verbal flubs.

In Stir’s case, his robot minder suggested he let his pitch rise and fall more to sound less monotone, something he says isn’t the easiest task, given the technical material he covers.

“It’s technology for corporate tax software,” Stir says. “No one’s going to carry me out of the room on their shoulders.”

Joseph Zalkin, 65, a retired emergency medical services worker in Raleigh, N.C., says the idea of recording and analyzing meetings is “Big Brother personified”—and extremely helpful. Zalkin sits on multiple committees associated with a foundation and a local university, and on days when he is double-booked, he will send in an AI-powered notetaker to silently listen in, transcribe and recap what was said. The follow-up reports hit his inbox from 20 minutes to two hours later.

To Zalkin’s amusement, the tools also offer punctilious summaries of small talk that a human notetaker might skip, including a recent discussion about an outpost of Bucee’s, the gas station and convenience store with a cult following, that opened nearby.

The reports Zalkin receives for meetings he attends also note whether he arrived on time and the number of times he interrupted people. During one recent family meeting about investments, the answer was 14 times—mostly to interject over his brother, Zalkin says. “I’m sure it was important to move things along,” he says.

His brother Andrew, 69, says he didn’t mind. “We all seem to talk a lot,” he says of the family, adding that he, too, is prone to interrupting.

Virtual meeting hosts usually have to click “accept” to let in the AI assistant, and with most tools, will display a notification for participants indicating that the meeting is being recorded. Still, some users say they have abandoned such technology after finding it too creepy.

“It’s like having a conversation with someone at a coffee shop, and you look out the window and all you see are a pair of eyes looking at you,” says Zack Schwartz, 33, the Chicago-based founder of a design firm who pulled the plug after experimenting with AI in meetings this spring.

The tools present tricky social questions, such as the etiquette around sending AI assistants to meetings without personally attending.

“There’s a weird power dynamic there,” says Jessica Malnik, a marketer in her 30s. Lately, when attending webinars in her field, Malnik says, she has noticed many more AI notetakers showing up. Each one occupies a faceless, darkened Brady Bunch square, which she says can create a weird feeling in a virtual room. And though users of such tools usually name them accordingly—“So-and-So’s Notetaker,” for example—they sometimes give more humanlike names that camouflage their presence, she says.

During one recent event, she saw a participant arrive right on the dot and remain without saying a word, even as others were chatting and saying hello. She recalls thinking: “This is obviously a bot.”

Link to the rest at The Wall Street Journal

PG is trying to decide whether an annoying person or an AI would be more damaging to the possibility that real work might get done.