These are the author's own words, unchanged. Mimansa editors picked the passages, and · · · marks where parts are left out. The full article adds the slow-down in effort, the scrutiny on careful colleagues, a recruitment example, and how HR can roll the rules out.
You're reading the full article · 4 min
Parts that are not in the key passages have a line beside them, like this one.
Kyun · your host
Today's question: is AI-made work making us trust each other less? Over to the author.
Fig. 1 · Looks finished
AI slop is everywhere on social media. Posts that say nothing. Images that look impressive until you look closely. Videos produced because someone can produce them, rather than because they have something to say.
Do you think this is limited to our personal lives?
It is happening at work too.
I recently spoke to a CEO who runs a Global Capability Centre (GCC) in India. Every week, his team prepares a report that he sends to senior leaders in the US. He was worried that his team might be using AI without reviewing the output properly. He did not want to send AI workslop under his name.
His question was simple: how could he tell whether the report was good without reading the whole thing carefully?
I could not give him a shortcut. That is the problem.
Kyun
So what does workslop look like on the page? The author starts with an example.
Question 1 of 5
What does AI workslop look like?
Imagine opening a report and reading this:
“We are leveraging our talent capabilities to drive transformational outcomes and create a more agile organisation.”
Which capabilities? What changed? What outcome? Has the team solved a problem, or simply written a sentence that sounds like work?
AI workslop uses polished language without saying anything concrete. Sentences are long and indirect. Observations could apply to any company. Recommendations say “continue to explore” without explaining what anyone should do next.
· · ·
The work can look more capable than the person’s understanding of it.
· · ·
Question 2 of 5
Why does this create mistrust?
When a colleague sends you a report, you assume they have thought through it. You may disagree with their conclusions, but you expect them to understand and explain their work.
AI makes that assumption harder to hold.
A strong report and a weak report can look almost identical. Both have clean slides, confident headings and fluent English. You discover the difference only when you read closely or ask questions.
· · ·
A recommendation sounds sensible until you ask how it would work. A number appears convincing until you ask where it came from. The person who sent the report cannot explain either.
· · ·
AI is changing how much we trust the work we receive from other people.
Kyun
The report took one hour to write. So where did the rest of the work go?
Question 3 of 5
Who pays for the time AI saves?
The writer finishes in one hour what previously took two days. But the manager now has to check the numbers, remove repetition, rewrite recommendations and ask what the team actually means.
· · ·
Are we increasing productivity, or saving the writer time by giving the reader more work?
Question 4 of 5
Will better AI skills solve this?
· · ·
But knowing how to use AI and being willing to check its work are different things.
· · ·
Mistrust will not disappear simply because AI becomes familiar. It will reduce when colleagues repeatedly receive work they can rely on.
Question 5 of 5
What can companies actually do?
HR could start with one recurring report that employees complain takes too long to read. Work with that team to introduce a few rules:
A 300-word weekly update: what changed, what needs attention and what decision is required. Supporting detail goes in an appendix.
Sources beside important numbers: the reader should not have to chase the sender for them.
A specific recommendation: what should happen next, who should do it and by when.
A short walkthrough when needed: ask the sender to explain the recommendation and the evidence behind it.
Kyun
Want to try these rules on one report this week? The kit has the templates.
When work fails these checks, send it back with specific questions. Managers
AI slop is everywhere on social media. Posts that say nothing. Images that look impressive until you look closely. Videos produced because someone can produce them, rather than because they have something to say.
Do you think this is limited to our personal lives?
It is happening at work too.
I recently spoke to a CEO who runs a Global Capability Centre (GCC) in India. Every week, his team prepares a report that he sends to senior leaders in the US. He was worried that his team might be using AI without reviewing the output properly. He did not want to send AI workslop under his name.
His question was simple: how could he tell whether the report was good without reading the whole thing carefully?
I could not give him a shortcut. That is the problem.
Kyun
So what does workslop look like on the page? The author starts with an example.
Question 1 of 5
What does AI workslop look like?
Imagine opening a report and reading this:
“We are leveraging our talent capabilities to drive transformational outcomes and create a more agile organisation.”
Which capabilities? What changed? What outcome? Has the team solved a problem, or simply written a sentence that sounds like work?
AI workslop uses polished language without saying anything concrete. Sentences are long and indirect. Observations could apply to any company. Recommendations say “continue to explore” without explaining what anyone should do next.
A report may have 20 slides, yet you finish without knowing what happened or what the team recommends.
Two years ago, preparing a presentation took human effort. Someone had to decide what to include, write the sentences and assemble the slides. Weak reports existed, but producing them still took time.
Now, someone who needed days can produce a presentation in an hour. The output looks better, but their understanding has not necessarily changed.
The work can look more capable than the person’s understanding of it.
There is another problem. Someone who previously put effort into a thoughtful presentation could stand out. Now everyone can produce polished slides.
If managers cannot distinguish the thinking behind those presentations, will the people who previously did excellent work keep putting in that effort? If an unchecked AI presentation gets the same appreciation, what are we encouraging?
Question 2 of 5
Why does this create mistrust?
When a colleague sends you a report, you assume they have thought through it. You may disagree with their conclusions, but you expect them to understand and explain their work.
AI makes that assumption harder to hold.
A strong report and a weak report can look almost identical. Both have clean slides, confident headings and fluent English. You discover the difference only when you read closely or ask questions.
A strong report and a weak report can look almost identical.
A recommendation sounds sensible until you ask how it would work. A number appears convincing until you ask where it came from. The person who sent the report cannot explain either.
After this happens a few times, you start questioning everyone’s work. Even careful colleagues face more scrutiny because their reports look similar.
AI is changing how much we trust the work we receive from other people.
तराज़ू · Fig. 2 · Who carries the weight
Kyun
The report took one hour to write. So where did the rest of the work go?
Question 3 of 5
Who pays for the time AI saves?
The writer finishes in one hour what previously took two days. But the manager now has to check the numbers, remove repetition, rewrite recommendations and ask what the team actually means.
The GCC CEO faces this every week. His team can produce the report faster, but he still has to establish whether it is worth sending under his name.
Recruitment has a similar problem. AI makes CVs sound impressive, leaving recruiters to find the actual experience behind the language. Some use AI to shortlist CVs that candidates used AI to write.
Are we increasing productivity, or saving the writer time by giving the reader more work?
Question 4 of 5
Will better AI skills solve this?
Some workslop comes from a capability gap. People do not yet know how to give AI enough context, question its answers or improve its output. Practice and training should help.
But knowing how to use AI and being willing to check its work are different things.
Someone can be excellent at prompting and still forward a report they have barely read. If that report gets accepted, there is little reason to change.
Mistrust will not disappear simply because AI becomes familiar. It will reduce when colleagues repeatedly receive work they can rely on.
Kyun
Skills help. Now the author turns to what a company can change.
Question 5 of 5
What can companies actually do?
HR could start with one recurring report that employees complain takes too long to read. Work with that team to introduce a few rules:
A 300-word weekly update: what changed, what needs attention and what decision is required. Supporting detail goes in an appendix.
Sources beside important numbers: the reader should not have to chase the sender for them.
A specific recommendation: what should happen next, who should do it and by when.
A short walkthrough when needed: ask the sender to explain the recommendation and the evidence behind it.
Kyun
Want to try these rules on one report this week? The kit has the templates.
These are examples, not limits for every type of work. A short report can still be useless.
HR can put these expectations into existing reporting templates and manager training. Show an actual vague update alongside a useful rewrite so employees can see the difference.
Fig. 3 · Sent back, with questions
Where the author leaves it
When work fails these checks, send it back with specific questions. Managers
Kyun
Now you. Where do you land?
Tap one to see how other readers answered.
Kyun
Those were the key passages. The full article stays longer on the cost to careful colleagues, and on how HR can bring in the rules.
Use this at work · Toolkit
Clear Update Kit
Write short updates your reader can trust, check and act on.
