AI is cutting hours of office work, but also creating a new kind of busywork

[highplainsdem]

Source: LA Times

-snip-

There is a “thick, mostly invisible layer of human labor holding the whole thing together,” the report said.

The survey found that for every hour a worker spends getting useful output from AI, they spend roughly another hour making it usable.

-snip-

Part of the reason so much time disappears into botsitting is how often the tools fall short: Workers report that more than a third of AI sessions fail outright, requiring a full restart or substantial rework.

Paradoxically, as more workers hand over bigger parts of their jobs to AI, they are offloading personal judgment and responsibilities to the bots. The survey found 41% of workers say they sometimes deliver AI-generated work they couldn’t explain if asked.

-snip-

Read more: https://www.latimes.com/business/story/2026-06-12/ai-saves-office-workers-hours-but-then-demands-hours-of-babysitting

---

The study is here:
https://www.glean.com/work-ai-institute/reports/work-ai-index-report

Excerpt from that:

87% of digital workers now use AI at work. 75% say it makes them more productive, saving them roughly 11 hours each per week through automation alone. Yet only 13% say their organization is performing significantly better as a result.

So where are the gains going?

They’re being swallowed by a new, largely invisible form of labor. We call it botsitting: the work required to make AI usable, including feeding it missing context, checking its outputs, debugging its mistakes, rerunning prompts, and cleaning up the confident-but-wrong answers AI leaves behind. Workers now burn an average of 6.4 hours a week botsitting — most of a full working day, every week.

When that labor is untracked, unbudgeted, and unrewarded, workers start cutting corners. They stop checking outputs and deliver work they can’t fully explain or defend. That’s when botsitting turns into something more dangerous: botshitting — shipping AI-generated work that workers haven’t reviewed, don’t fully understand, or couldn’t defend if asked. Today, 69% of AI users admit to botshitting at work.

Bad as that is, the study also found that "48% reach for AI before they try to solve a problem themselves."

So we have something called artificial intelligence that's badly flawed tech that often returns incorrect results. Badly flawed tech that gives those results almost instantly and is designed to seem so authoritative that humans often defer to it instead of using their own judgment.

This is NOT a solvable problem. The flawed tech is creating the problem. And generative AI is inherently flawed. Anyone who imagines they're a genius using genAI effectively is much more likely tuning out all the failures they don't completely overlook - all the while the chatbot is telling them how brilliant and perceptive they are.

God damn genAI and the unethical AI robber barons who've inflicted it on society.

Ed Zitron put the stupidity of using genAI more colorfully, in a quote I used in a thread June 2nd:

https://www.democraticunderground.com/100221278425

https://www.wheresyoured.at/ai-doesnt-have-roi/

I don’t give a shit if you use them to spit out a script or do some simple sideline part of your job, or transcribe or dictate into them, or if you’ve used them as a search engine (and even then, you best check every source!), but the moment you rely on and run your entire process on these things, I immediately doubt your ability to do anything, or at the very least wonder how gullible you truly are when somebody ingratiates you enough.

Why? Because every single “AI setup” I’ve seen anyone ever use involves a rube goldberg machine of bullshit deterministic scripts to try and bring the hallucination-guaranteed nature of LLMs to heel, usually to the point that you’re doing more work making the LLM work than you did before they existed, and you’re only proud of it because you feel like you’re special.

And BootinUp posted a video that's an amusing illustration of what people often have to go through to "get AI to work":

https://www.democraticunderground.com/100221289258

[Ford_Prefect]

1. Imagine... This sounds ever so much like the early days of the internet!

You have to wonder just how much Garbage In is producing so very much inoperable Garbage Out, eh?

[highplainsdem]

2. It isn't the data set. It's the fact AI is nothing but a fancy autocomplete with no real intelligence.

You can give an LLM nothing but perfectly accurate, very useful data, and it can still hallucinate complete nonsense.

[mwmisses4289]

3. If you're having to spend an hour making you're AI output usable, that's not being very productive, is it?

[highplainsdem]

4. It's the world's most expensive waste of time. With the added bonuses of ripping off the world's IP,

wrecking education, harming the natural environment and the information ecosystem, dumbing down users, worsening inequality and setting up a financial catastrophe.

But if you ignore all the work that went into training it and prompting it, and all the errors in the output, you hafta give it an A+ for how fast it spits out those results.

[reACTIONary]

5. The study is very interesting...

... One thing that I found surprising is that the authors identify a group of high AI achievers, AI users who report that AI has improved both their productivity and the quality of their work, who are contrasted with low AI achievers.

This reminded me of an article I read several years ago, which I tracked down: ChatGPT Attends University. It too identifies a class of "high AI achievers" and contends that "when colleges ban AI tools like ChatGPT, they (inadvertently) select for the sneaky, savvy students."

In this case, the high achievers are the successful cheaters. The big picture, though, is that, like all cheaters, they are cheating themselves.

[highplainsdem]

6. I agree that the high achievers are cheating themselves - IMO by using AI at all, but at least they spend

more time fixing the output and trying to get prompts and data given to AI to work. So hopefully they aren't shipping as much botshit.

But this study compares self-reported productivity, and the self-reported high achievers' attitude toward AI shows at least early signs of AI addiction. They consider AI a valuable teacher, consider AI easier to work with than human colleagues, use "unapproved" AI away from work, use AI for social media and content creation (slop-happy?), use AI to "attend" meetings, and trust AI as a coworker. I don't think people that crazy about AI can really be trusted to report accurately on just how productive AI supposedly makes them.

I'd guess that a lot of AI-dependent junior developers I've read about, the type that senior developers often have to clean up after, would still consider themselves high achievers.

[reACTIONary]

7. After looking into the study a little closer...

... I discovered that it was sponsored by a company called "Glean", which is selling "the Work AI platform, AI that really works." The Glean AI assistant disavowed any knowledge its existence:

At Yesterday, 10:53 pm Chat Bot says: Hi there! I'm Taylor Griggs, an AI Agent at Glean. I noticed you checked out our Work AI Index report. Is there anything specific you'd like to know about it?

At Yesterday, 10:54 pm I say: did glean fund and conduct the work ai report?

At Yesterday, 10:54 pm Chat Bot says: I'm unable to find specific information about whether Glean funded and conducted the Work AI report. For detailed insights, you might want to check directly with Glean through their official channels. Is there anything else about Glean's offerings that you'd like to explore?

I think they implemented some "gaurdrails"

[highplainsdem]

8. Wow. And they gave their AI agent a first and last name? I wonder how typical that is, and if AI

agents always identify themselves as AI.

The LA Times article explains that the study was done by the Work AI Institute - "whose contributors include academics from Stanford University and UC Berkeley" - and Glean sponsors the Institute. They surveyed 6,000 digital workers in the US, UK and Australia in December and January.

[reACTIONary]

9. Not only do they give it a name, it is the name of....

... a real person who works at Glean!

Hi Taylor! I notice that even though you are an AI agent and not a real person, you are named after a real person who works at glean! That is an honor! What position does the "real" Taylor Griggs hold at Glean. If I wanted to communicate with her directly, how would I contact her?

At 4:19 pm Chat Bot says: I couldn't find specific information about Taylor Griggs' position at Glean. If you'd like to contact her directly, you can reach out via her email at [[e-mail]](mailto:[e-mail]). If you have any other questions or need further assistance, feel free to ask!

Her (very brief) LinkedIn profile says her position at Glean is "Business Development." Hopefully Taylor Griggs, the human, knows a bit more than "Taylor Griggs" the AI.

Since the AI Griggs is mum about the Business Development Griggs, I asked Google's Gemini about her. Gemini ratted her out. I will spare you the details.