FIR #525: The Consequences Businesses Will Face from Work Slop

We have been hearing about work slop for a while now, and evidence is mounting that it can create long-term problems for companies. Yet we are hearing very little (if anything) about what businesses are doing about it. There’s new research, which Neville and Shel discuss in this episode, along with some thoughts about actions companies can take.

Links from this episode:

• Why People Create AI “Workslop”—and How to Stop It
• Navigating AI in the Workplace: 2026
• Harvard Business Review warns AI ‘workslop’ is rotting companies from the inside
• Botsitting, botshitting, and the hidden human labor of AI at work
• 27 Seconds to Breach: The AI “Workslop” Tax Hitting 2026 Bottom Lines
• How to Eradicate AI ‘Workslop’ From Your Office

The next monthly, long-form episode of FIR is tentatively scheduled to drop on Monday, August 24.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email fircomments@gmail.com.

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript:

Shel Holtz

Hi, everybody, and welcome to episode number 525 of For Immediate Release. I’m Shel Holtz.

Neville Hobson

And I’m Neville Hobson.

Shel Holtz

And Neville, it is great to have you back on the show. You’re well enough to record again.

Neville Hobson

Yeah, it’s certainly been an interesting few weeks, including the kind of disconnection that I did from just about literally everything with this heat exhaustion that I suffered, which was quite profound and far more serious than I thought, according to the doctor.

And then the R&R, as I described it—the rest and recuperation—slotted in nicely with our existing family plans here to spend some time in Novato, Marin County, in California, which we did last week. And as some listeners will know, you and I and the family, in fact, all met together, had lunch in a delightful restaurant in San Francisco last Thursday, and took some pictures, which is what people saw.

So I’m back in the UK now. It was a worthwhile visit, and, well, here we are doing FIR again. It feels like déjà vu all over again. That’s great.

Shel Holtz

It is back to normal. I should tell people that if you’re curious about the wonderful restaurant we ate at in San Francisco, it’s called Waterbar, and my company, Webcor, built the building that it’s in. So it was fun to have you there and to say, “We built this.”

Neville Hobson

Yeah, it was. I agree. The hospitality was really good in that restaurant. Very nice service, friendly people, excellent food, of course. It’s focused on fish, so if you’re a fish connoisseur, this is the place to go in San Francisco, I would say. Yeah, really nice.

Shel Holtz

Yeah, and right on the water. You get a view of the Bay Bridge and a good part of San Francisco Bay. So, yeah, it was great to see you in person. It’s been since 2019.

Neville Hobson

Yeah, a long time.

Shel Holtz

And here we are looking at each other over Riverside, but that’s not the same.

Neville Hobson

No, the next best thing, but not the same.

Shel Holtz

Yeah, being together in real life really matters. But today’s topic is one that we were planning to discuss just before you took ill, so we’re going to return to it.

Neville Hobson

Yeah.

Shel Holtz

We’re going to return to it.

Neville Hobson

We are. And this is based on an article that caught my eye a few weeks ago in the Harvard Business Review. It was actually published back in January, but having returned to it, I think it’s even more relevant than it was then.

The title of the article is “Why People Create AI ‘Workslop’—and How to Stop It.” Timely topic, I’d say. The authors use the term “workslop” to describe low-effort, AI-generated work that looks polished on the surface but ends up shifting the real work onto whoever receives it.

We’ve probably all seen examples: reports full of generic language, presentations that say a great deal without saying very much, emails that are technically fine but leave you wondering what the sender actually thinks.

Since that article appeared, the wider idea of AI slop has become much more prominent. We’re hearing the term applied to the flood of cheap, AI-generated material appearing on social networks, in publishing, marketing, and elsewhere. There’s even research suggesting that people increasingly use “AI slop” as a judgment about authenticity. Something doesn’t necessarily have to be AI-generated to be dismissed as slop; it just has to feel like it.

Perhaps the most useful new evidence comes from research by the Society for Human Resource Management, SHRM, the world’s largest professional association dedicated to the practice of human resources management. Published in June, SHRM’s Navigating AI in the Workplace: 2026, based on more than 5,000 U.S. workers, found that 41% use AI at work, and 44% of those users describe at least some of their output as AI slop. Early-career employees feel the greatest pressure to adopt AI. SHRM also found greater engagement and commitment where organizations take an open approach to AI integration.

What interested me most about the HBR article was its argument about why workslop happens inside organizations. The authors suggest that it isn’t fundamentally an AI problem at all. It’s a leadership problem.

Many organizations are telling employees simply to use AI without defining what success looks like, without giving people sufficient guidance or confidence, and without creating an environment where it’s safe to experiment, ask questions, or admit uncertainty.

In that environment, people can end up optimizing for appearing to use AI rather than actually using it well. Putting it simply, the article says it’s the result of unclear AI mandates and overwhelmed teams. Leaders are issuing vague directives for employees to start using extremely powerful tools, while many of those employees are overburdened, psychologically depleted, and operating in environments where it doesn’t feel safe to admit uncertainty or ask for help.

Evidence published since the HBR article seems to reinforce rather than undermine that argument. We’re seeing more research showing significant numbers of employees encountering poor-quality, AI-generated work while, at the same time, employees are under increasing pressure to demonstrate that they’re using AI.

There’s another dimension to this that I think is particularly interesting: productivity.

Imagine that something that previously took me two hours to produce now takes me 20 minutes with AI. That’s a significant productivity gain for me. But suppose I send it to Shel—to you—and you then spend an hour trying to understand what I mean, checking my claims, and correcting some mistakes that I hadn’t spotted. Have we actually improved productivity, or have I simply transferred the cost of my work to Shel?

That’s essentially what workslop does. And some recent research into AI-assisted work describes this as a kind of tragedy of the commons. Individuals can appear more productive while imposing additional costs on everybody else.

If an organization measures AI adoption, output volume, or individual efficiency rather than the performance of the whole system, workslop might actually look like successful AI transformation.

That struck a chord because it echoes something we’ve discussed on FIR more than once. Whenever we hear stories about AI failing in organizations, the technology itself often isn’t the real issue. AI has a remarkable ability to expose weaknesses that were already there: unclear leadership, poor communication, lack of trust, unrealistic expectations, and organizations measuring activity rather than outcomes.

And there’s an important question of individual responsibility here, too. Poor leadership may create the conditions for workslop, but that doesn’t absolve the person who sends it. If I knowingly send something I haven’t properly considered or checked simply because an AI produced it, that’s also a failure of professional judgment.

So perhaps workslop isn’t really the diagnosis at all. Perhaps it’s a symptom.

Is AI creating a new workplace problem, or is it simply shining a very bright light on old management problems? What happens when leaders measure AI adoption instead of better decisions, better collaboration, and better business outcomes? And as AI makes producing apparently polished work almost effortless, does human judgment actually become more important rather than less?

The Harvard Business Review article may now be seven months old, but the questions it raises are becoming more urgent, not less. Shel?

Shel Holtz

Without question. And there—

Neville Hobson

Mm.

Shel Holtz

—there is newer research. I found work done by the Work AI Institute. They did a 2026 research project called The Work AI Index. And in addition to workslop, they have a new term in there called “botshitting.” I kid you not.

Neville Hobson

Ha ha ha.

Shel Holtz

Botshitting describes employees knowingly shipping AI output they believe is wrong. And 12% of employees who were surveyed for this research admit to doing this. Interestingly, when they are basically caught, they blame AI. They say, “The AI did this.” Forty percent of workers blame the AI rather than themselves.

There was also a study in January 2026 from Workday. Some 3,200 leaders found that 37% of AI productivity gains are immediately lost to rework because the workslop isn’t adequate, with employees spending an average of six hours a week correcting or rewriting flawed AI content.

Built In reported workslop now affects roughly 40% of employees and costs about $186 per worker per month. And there was a more recent study—BetterUp sourced this figure—that found that 66% of workers are spending six-plus hours every week fixing AI errors. So this is serious stuff.

And by the—

Neville Hobson

Mm.

Shel Holtz

—way, nearly every recent report on workslop ties this to layoffs associated with AI. Workers and outlets are connecting 2026 tech job cuts to premature AI-driven headcount reduction. And the workers who are left behind are absorbing that workslop rework burden. So we cut people to use AI; now the survivors are—

Neville Hobson

Mm.

Shel Holtz

—cleaning up after it. That’s an important framing to look at.

And one last thing I’ll mention from some research—this was just from June of this year. They found that the errors in workslop go beyond the work that somebody else needs to do to correct it. It moves downstream through teams and into the organization’s collective knowledge base. And that knowledge base just deteriorates.

So it’s not just individual wasted hours, but this slow erosion of what the organization actually knows to be true. This is a very serious problem, and I don’t see a lot being said about measures that organizations are taking to address it. I think some measures need to be taken pretty soon, or this is going to get out of control.

Neville Hobson

Yeah. The Harvard Business Review article does have some suggestions about what organizations can do to address it, which I’ll mention in a minute. But they also include some examples. I found these rather interesting. Not a whole lot—three only—but nevertheless, they talk about the toxic effect workslop can have on workplace dynamics, breeding mistrust and leading team members to think less of the sender’s intelligence and trustworthiness, among other traits.

They talk about their ongoing research, where they’ve heard a number of examples of workslop seeding ill will, eroding trust, and generally having a corrosive effect on workplace morale.

There’s one where I thought, I can imagine this. One example they give is an employee at a technology company who told the Harvard Business Review that he’d noticed the tone in his performance review was unlike his manager, and that the document recycled content from his self-evaluation. The experience made him feel unvalued and underappreciated, and he gave up all hope that he would ever be promoted.

So, reading between the lines, what happened? Whoever did that grabbed some of the employee’s self-evaluation and had an AI produce the review. I bet that’s what happened there.

We don’t know the details behind that, but that’s a really good example of what you could see happening in an environment where there is a lack of clarity, all those things the Harvard Business Review points out, and the pressure to deliver in some form or another.

So, for instance, doing an employee evaluation would come into that area where the manager has cut corners and used AI to do it. And I could imagine that is happening a lot.

And in relation to the examples that the Harvard Business Review piece includes, they say something quite interesting. Many of the responses focused on the productivity costs—the time people wasted dealing with each instance of workslop that crossed their desks. But they note that what should really worry leaders is the impact workslop can have on human relationships.

And I think that’s a point to hammer home. It’s a kind of soft thing. You don’t necessarily see it, but it’s there. And some of these examples make that very clear indeed. So that’s a manifestation of the management failure behind all of this, Shel, don’t you think?

Shel Holtz

I do. And I think it’s because we’ve been rushing headlong into this rather than pausing to strategize it.

And I think we’ve talked about that deterioration of human relationships once before. The situation is that people are not picking up the phone or sending an email off to the local subject-matter expert because Claude has the answer. ChatGPT has the answer. So you’re getting less of that interaction within the organization.

This is causing people to assume some of the work that is outside their area of expertise, outside their lane, if you will, instead of reaching out to the people in the organization who actually have that expertise so they can do the work that previously was part of their job.

Another interesting thing you noted was that somebody felt this was leading to a lack of promotion opportunity.

Neville Hobson

Mm.

Shel Holtz

I’m reading a book right now, an excellent book. We were talking about it before we started recording. It’s called Robot-Proof. And one of the things she talks about is this economic phenomenon of deprofessionalization, where AI takes on a certain amount of the work, and therefore you, as the professional, don’t need to know or do as much as you did before.

And it lowers your overall value. It pushes the wage down for people who are doing that. The author, Vivienne Ming, uses an analogy of Jiffy Lube, which in the U.S. is a place where people take their cars to have their oil changed.

And she says, “What if colonoscopies worked that way?” You don’t need a doctor to do a lot of this work because a lot of the work in a colonoscopy is pretty routine and AI can do that. Now you need just the Jiffy Lube guy there to take the hose and do—

Neville Hobson

Yeah.

Shel Holtz

—the insertion, looking at the camera to make sure it’s right. But that’s more of a technical skill than a deep medical skill that is taught in medical school and then honed over years of experience and practice.

And this is the road I think we’re headed down if we don’t take some steps to fix it. The people who were brought into an organization based on education and years of experience and subject-matter expertise, and perhaps even thought leadership, no longer need all that stuff. They just need to do the base-level effort that is required by a human following the instructions of the AI.

So I think organizations need to look at this very seriously: how this work gets done, what is required of the outputs, and having guardrails in place to make sure that people aren’t skirting around it.

Because let’s face it, if the cost of producing this stuff continues to drop, there’s an incentive on the part of the organization to go ahead and use this because the cost is lower. It’s hard to see the impact of that on sales and reputation in the short term. In the long term, I think it absolutely will have that impact.

Neville Hobson

Yeah, I’d say you’re right. I mean, one thing that I’ve been talking about all year in everything I’ve written or said about artificial intelligence in organizations is that it’s about the people, not the technology.

If we’re implementing AI or doing a rollout or planning it, or telling people, “Use AI,” as a leader, you have the responsibility to make that as crystal clear as possible and focus on the people.

So one of the things I did like about the Harvard piece was the concluding remarks they make, starting with this: “The greatest irony of all is, to make AI work at work, we need to get better at being human.”

Absolutely spot on.

Leaders need to make space for the unpolished, slower but more rewarding work of human collaboration. Without organizational changes that enable agency and trust rather than AI mandates for overburdened teams, we’ll all drown in the sludge of workslop.

A very well-put concluding statement, I think. But the serious element of that is something that, in my opinion, I’m amazed that leaders don’t seem to get at all.

They talk about efficiencies. They talk about cost savings. They talk about all these things, and yet there’s nothing about: How do we help employees become more productive themselves and develop themselves, too? How do we help them, in a sense, augment their skills using AI?

Now, that’s not to say—and I’m not suggesting for a second—that AI leaders generally just don’t think about that at all. I believe they do, but it’s way down the priority list in terms of how they communicate this. They need to get that to the top of the priority list.

But more than just talk about it, they need to put these things in place. And there are many things they can do. There’s, in a sense, rebuilding trust with people.

We’ve talked about examples in recent episodes of layoffs that are very clear to see. People are talking about them all the time, and that is happening. We talked about a really great one a few episodes back about what Ford Motor Company is doing with the graybeards. Lovely name. We’ll pass on the age—

Shel Holtz

Mm-hmm.

Neville Hobson

—thing on that. But it’s about the organizational collective memory people have of how things are done and what worked in the past, among people who were laid off and are gone.

So they rehired a whole bunch, and they’ve benefited hugely, as the articles we cited go into in some detail, including, I believe, if I recall, Shel, they were the number-one automaker for a particular reason that they attribute directly to rehiring all these graybeards.

So that is a great way of doing it. But it’s just one example. There are undoubtedly more fundamental things organizationally that people can do to rebuild this connection between the organization and the people who make up that organization—the employees, mostly.

So there’s a lot of work to be done. And yet none of this is rocket science to leaders at all, I don’t believe. Just because we’ve got this newfangled technology on the scene, you’re telling people to start using it without giving them the proper guidance.

So these, to me, are Leadership 101 things. And if no one’s doing this, then communicators, let’s step up to the plate and help. That’s what I think we should be doing.

Shel Holtz

Yeah, absolutely. And I have been reading more and more research that talks about what happens when you delegate your thinking to AI.

Rather than doing the research yourself and the writing yourself, which helps you think through the issue, it actually has an impact on your ability to think and to learn. So there are a lot of reasons I think that we need to take this seriously.

I think the first thing communicators need to do inside their organizations is raise the alarm. Because I think this may be an issue where everybody’s heard “slop” and everybody’s heard “workslop,” but I don’t know that everybody in the organization—especially the decision-makers—is up to speed on what the research is saying about all of this and the deleterious impact it can have on our organizations and the outputs that we produce.

So, yeah, I think sharing this data—and we will have links to the research in the show notes—and making the people who make the decisions aware of what this could mean to the future reputation and earnings of the organization is the place to start.

Neville Hobson

Yeah, I agree.

So, simply, there are good tips in this discussion. The links to the articles—please read them. They’re really, really useful. You’ll find a lot of insights in all the ones that we’ll have in the show notes. So there you’ve got your route map, I would say.

Shel Holtz

And that’ll be a 30 for this episode of For Immediate Release.

The post FIR #525: The Consequences Businesses Will Face from Work Slop appeared first on FIR Podcast Network.