Editorial
Workslop. That’s the term researchers at Stanford and BetterUp use to describe the surge of AI-generated content that looks like “good work” but “doesn’t actually advance the work itself.” While their survey of 1,150 — and counting — U.S. employees has drawn fair criticism for its methodological woes, it struck a nerve because it captured what employees are increasingly tripping over every day.
Workslop isn’t new. It’s just the latest strain of productivity theater — the workplace ritual of looking busy instead of creating value. Roughly two-thirds (65%) of workers admit to doing it. At last month’s Ai4 conference, I moderated a panel with Andy Ballester, co-founder of GoFundMe and now EyePop.ai, and Cyril Gorlla, co-founder and CEO of CTGT. Both warned about the rise of what they called “AI slop”: cheap, auto-generated code churned out through “vibe coding.” The code looks solid at first glance, maybe even runs fine in a demo, but buckles under real-world scale and production.
AI slop is now infecting the workplace’s most notorious time sink: meetings. For decades, they’ve been the poster child for productivity theater — calendar clutter dressed up as busyness. Now AI has spawned a new strain: “meetingslop.”
You’ll often spot meetingslop in four familiar strains:
1. Meeting Assbots: Bots That Bulldoze Your Calendar
One of the most common strains of meetingslop is what Richard Clayton proposes calling the “assbot”: scheduling bots that shove meetings onto your calendar with all the subtlety of a battering ram.
A couple of months ago, as part of our latest research at Glean’s Work AI Institute, Stanford Professor Bob Sutton and I tried to schedule a meeting with someone whose assistant was Kiran, an “intelligent” scheduling bot. Kiran kept prodding Bob to confirm a time while he was on vacation. His out-of-office reply was on, but it didn’t matter. As Bob later recounted, Kiran kept hammering both of us with requests. I even emailed it privately to say it should just book it, hoping it’d stop. But Kiran wasn’t intelligent enough to deviate from its script. The pings kept coming.
Assbots don’t just waste time — they make their owners look bad. As UC Santa Barbara professor Paul Leonardi has found, the blame rarely lands on the bot. It usually falls on you, the human who trusted it, while the bot struts away with a false badge of productivity.
2. Meeting Botshit: AI That Hallucinates in Meetings
The second strain of meetingslop is what Tim Hannigan, professor at the University of Alberta, calls “botshit”: large language model hallucinations that humans accept uncritically.
In meetings, it shows up as AI notetakers inventing phantom action items, misattributing decisions or confidently recording things no one actually said. OpenAI’s Whisper, an AI speech-to-text system once marketed as approaching “human-level robustness,” has been shown to fabricate words, sentences and even entire narratives. A University of Michigan researcher reported false content in 80% of public meeting transcripts he examined.
The best defense against botshit is grounding AI in real work context. When it understands who does what, how work gets approved, what’s already been discussed and where decisions are logged, it’s far less likely to spin fiction.
3. Meeting Botspam: AI That Makes It Harder to Find the Real Humans
The third strain of meetingslop is what I call “botspam.” It happens when multiple participants bring their own AI notetakers into the same meeting.
When humans log on and see that half the attendees are note-taking bots, the focus shifts from the discussion to the circus of machines transcribing it. One person’s Otter, another’s Fireflies, a third’s Zoom assistant — so many bots jockeying for notes that the meeting itself gets lost.
And when the meeting ends, instead of one clear record, you leave with five slightly competing versions of reality — each close enough to sound plausible, but different enough to cause confusion.
4. Meeting Botcop-out: AI That Lets You Phone It In
The darkest strain of meetingslop is what I call the “botcop-out”: when people stop showing up altogether and send their AI “digital twin” instead.
SaaStr founder Jason Lemkin recently shared that his Monday staff meetings dwindled from 16 humans to just three, along with 11 bots. Even if a bot can represent you, people notice when you couldn’t be bothered to show up. It signals the meeting — and, by extension, their time — doesn’t matter. And once colleagues decide you’ve checked out, trust erodes — and no transcript can repair the damage.
Stop Meetingslop Before It Spreads
Meetings have been dysfunctional for decades. As I unpack in my upcoming book, "Your Best Meeting Ever," if we’re not careful, meetingslop will pull them deeper into the swamp.
Start by setting ground rules for when AI belongs in a meeting. When should it take notes, set the agenda or stand in as someone’s digital twin? Each use requires a different level of consent, visibility and human oversight. Consider setting a simple rule: at least 70% of every meeting’s contributors must be human. If the ratio dips below that, cancel or reformat the meeting.
Next, run a role audit. If an attendee (human or bot) doesn’t have a clear role, they don’t belong. Use the “three-word test”: if you can’t sum up someone’s role in three words (“provide industry expertise,” “represent customer needs,” “transcribe the meeting”), they shouldn’t be there. And if multiple people or bots share the same three-word role, that’s a sign your meeting is bloated. Attendance should be earned, automatic. Just because a bot is easy to invite doesn’t mean it has earned the right to show up.
Then, close the loop. Make sure to collect feedback. Don’t just ask if the AI made the meeting more efficient. Ask if it changed how people behave. Did they show up better prepared, make decisions faster, or follow through on what they committed to in the meeting? Or did it make them more hesitant to speak up “on the record” comments, or sit back waiting for the bot to think for them?
Conference
Oct
20
Gartner IT Symposium/Xpo Orlando 2025
Register
Conference
Oct
27
Gartner HR Symposium/Xpo Orlando 2025
Register
AI lowers the bar for calling a bad meeting and can just as easily raise the cost of enduring one. If we’re not careful, we’ll keep piling on the meetingslop and be left with meetings so hollow that only the bots bother to show up.
Editor's Note: Learn more about other ways to improve dysfunctional meetings:
- What Are Meetings Costing You? — Workplace productivity, for one.
- Virtual Meeting Makeover: From Good to Great With a Few Simple Tweaks — With self-reflection, the ability to take and apply feedback and the willingness to make changes, you can transform your virtual meetings from good to great.
- Digital Twins in Meetings? Not Any Time Soon — Zoom CEO Eric Yuan recently suggested digital twins would attend meetings in our place. For now, that's more dream than reality.
Learn how you can join our contributor community.
