Written by Tim Green, narrated by AI. Listen to the full episode here.
🎙️ Season 1, Episode 24 | Duration: 20:13
Six thousand office workers across the US, UK and Australia were asked how much time AI saves them. They said eleven hours a week. Then they were asked how much time they spend feeding context, checking outputs, correcting errors and cleaning up afterwards. The answer was 6.4 hours. Roughly six of every ten hours AI appears to save are consumed by the labour of making AI work. The survey, published as the Work AI Index by Glean's Work AI Institute with researchers from Stanford, Notre Dame, Emory, UC Berkeley and UCL, gave that residual a name: botsitting.
This episode uses AI voice narration from ElevenLabs Studio.
The Arithmetic That Stops Working
A task used to take twenty minutes. Now it takes five minutes of generation plus fifteen minutes of checking. On the stopwatch, nothing has changed. On the productivity dashboard, everything has. The dashboard records five minutes of task completion. The fifteen minutes afterwards are logged as the worker doing their job. The tool has not saved fifteen minutes. It has reclassified them.
The Perception Gap
Only thirteen per cent of organisations surveyed said AI had significantly improved performance, against eighty-seven per cent of workers using it constantly. That gap is where the botsitting hours have gone. They have not disappeared. They have been absorbed into a category of labour the accounting system cannot see.
What Happens When Supervision Becomes Unaffordable
Sixty-nine per cent admitted to what the report calls botshitting: shipping output they had not verified, did not fully understand, or could not confidently stand behind. Forty-one per cent had sent work they would be unable to explain if questioned. Workers who reported botshitting were 3.8 times more likely to be looking for another job. This is not laziness. It is triage under an unfunded mandate.
The Downstream Cost
BetterUp Labs and the Stanford Social Media Lab surveyed 1,004 desk workers about workslop: output with the appearance of good work but lacking the substance. Forty per cent had received it in the preceding month. Each incident took an average of one hour and fifty-one minutes to sort out, roughly twenty minutes longer than doing the work properly would have taken. That came to roughly 186 dollars per employee per month. Workslop is botsitting skipped upstream, landing unpriced on somebody downstream.
The Forty-Year-Old Manual for This Problem
Lisanne Bainbridge published "Ironies of Automation" in 1983. Her central observation: the designer regards the human operator as unreliable and tries to design them out, but cannot automate everything. What remains for the human is the residue that could not be specified: the awkward, judgement-heavy fragments. The operator is left with the hardest parts, stripped of the easier parts that kept their skills sharp. Substitute a language model for a distributed control system and the paper reads like a memo from last week.
Why Checking Is Harder Than Doing
A junior colleague who does not know something produces work that signals its own uncertainty. A language model that does not know something produces work that is fluent, confident and internally consistent. The error sits in a sentence that looks exactly like every true sentence around it. The cost of finding it is not scanning for anomalies. It is independently establishing the truth of each load-bearing claim, which in the limit is the cost of having done the work yourself.
Key Sources
- Work AI Index - Glean Work AI Institute / Stanford / Notre Dame / Emory / UC Berkeley / UCL
- Ironies of Automation - Lisanne Bainbridge, Automatica, 1983
- Complacency and Bias in Human Use of Automation - Parasuraman & Manzey, Human Factors, 2010
- Workslop: The Hidden Cost of AI-Generated Busywork - BetterUp Labs / Stanford Social Media Lab
Listen to the Full Episode
🎧 Unveiling the Hidden Labor of Botsitting in AI Workplaces | Duration: 20:13
Subscribe on Apple Podcasts, Spotify, or your favourite app.
SmarterArticles is written by Tim Green, narrated by AI via ElevenLabs Studio. New episodes every Monday. Follow @humanin_theloop for updates.
Top comments (0)