ai-pocalypse Workers are getting lazy about using AI to do their jobs for them, and the results are both costly and increasing distrust in the workplace.

An ongoing study by Stanford’s Social Media Lab and behavioral research business BetterUp Labs says 40 percent of US workers are reporting AI-generated garbage, dubbed “workslop,” coming into their work lives in the last month. Essentially, staffers are sending around AI-generated material that may look impressive, but contains very little in the way of actionable facts and figures, which someone else then needs to sort out and turn into something useful.

Workslop is the machine-learning equivalent of spam, and the study claims that the amount of work involved in sorting out the wheat from the chaff, or facts from hallucinations, in such content, costs around $186 per employee per month in lost productivity, a few bucks less than a ChatGPT Pro account.

Furthermore, once someone receives this kind of content, over half of the respondents reported feeling annoyed, over a third said they were confused, and nearly a quarter said the messages offended them.

Also, recipients see senders themselves as more untrustworthy. Forty-two percent of survey respondents said that after receiving such data garbage, they trusted the coworker sending it less, and over a third took it as a sign that the sender was less creative and intelligent than they first thought. Many said such material was more trouble than it’s worth.

“It was just a little confusing to understand what was actually going on in the email and what he actually meant to say,” one tech boss told the researchers. “It probably took an hour or two of time just to congregate [sic] everybody and repeat the information in a clear and concise way.”

Unsurprisingly, the tech industry is one of the biggest generators of workslop, with professional services also highlighted as a key generator.

This stream of AI-generated effluent flows both ways, the survey found. Workers sent 18 percent of workslop directly to managers, but respondents said that 16 percent of such content came from managers themselves. It seems like both sides of the corporate world are happy to let AI do their thinking for them.

“It created a situation where I had to decide whether I would rewrite it myself, make him rewrite it, or just call it good enough,” one finance industry respondent told the surveyors. “It is furthering the agenda of creating a mentally lazy, slow-thinking society that will become wholly dependent [sic] upon outside forces.”

With more companies insisting staff rely more on AI – or face losing their jobs – the rise in workslop is somewhat understandable. Since staff are growing increasingly likely to use the technology, the temptation to take shortcuts is more probable and – like AI outputs in general, OpenAI has acknowledged – it’s better to put something out there than nothing at all.

But the study raises more questions about the actual productivity gains from AI in the workplace. A recent study by the UK government found no clear productivity improvement from introducing Microsoft 365 Copilot in the Department for Business and Trade, and research [PDF] by MIT reports that about 95 percent of organizations see no measurable return on investment from generative AI efforts.

But AI is everywhere, and firms are keen to make this happen. The survey is ongoing, but the early results aren’t good. While it’s easy to use AI to produce work that appears to be good enough, actually getting things right takes some skill. ®