There is a growing chorus complaining about “AI slop” in IT service management (ITSM).
Generic articles. Confident summaries that flatten nuance. Strategy outlines that look impressive until you try to implement them. The frustration is understandable.
But before we treat this as a technological failure, we should pause.
AI didn’t invent slop; it learned from us
If AI can produce slop at scale, the more uncomfortable question is not whether machines are degrading standards. It is whether those standards were already compromised.
AI was trained on our documents – our frameworks, our presentation decks, our policy papers, our certification manuals. If slop is everywhere today, we should examine its lineage before assigning its parentage.
What Do We Really Mean by “AI Slop”?
Slop isn’t simply bad writing. And it isn’t defined by whether a human or a machine produced it.
Authority Without Accountability
Slop is output that sounds authoritative but hasn’t earned that authority. It often contains pieces of truth. It may reference respected models. It may even be technically accurate in places. But it lacks grounding. It doesn’t surface its assumptions. It avoids the constraints that determine whether ideas work in practice.
Why Partial Truth Scales So Easily
Slop travels well because it isn’t tied too tightly to reality. It can be reused, reworded, and repackaged across sectors with minimal resistance. It feels safe because it avoids sharp edges. It sounds credible because it borrows familiar language.
It isn’t fabrication.
It’s partial truth, delivered with more confidence than context.
AI as Mirror, Not Origin
AI has undeniably accelerated slop production.
What once required time and effort can now be generated instantly. The speed is new. The scale is new.
But speed is not origin.
AI did not decide that transferable authority was desirable. It absorbed what we produced – the frameworks that travelled comfortably between industries, the “best practice” language that sounded right in almost any boardroom.
If abstraction was common in what we wrote, taught, certified, and sold, it shouldn’t surprise us that a machine trained on that material can reproduce it fluently.
When we blame AI for the surge in slop, we may be reacting to the visibility. What feels alarming is the volume.
The pattern itself may be older.
The Human in the Loop Is Not a Quality Guarantee
We speak about the “human in the loop” as though the phrase guarantees quality.
It doesn’t.
If the human is applying judgment – questioning assumptions, adding context, challenging weak reasoning – AI becomes an accelerator of depth.
If the human is skimming, polishing, and publishing because the output “sounds about right,” AI becomes an accelerator of slop.
The uncomfortable truth is that AI rarely exceeds the cognitive ceiling of the person guiding it. It reflects their scrutiny. It mirrors their tolerance for vagueness. It amplifies blind spots as efficiently as strengths.
A weak practitioner with a powerful model does not produce strong thinking.
They produce faster abstraction.
And when that abstraction is deployed – in strategy, governance, automation design – the machine is not accountable.
We are.
Is AI Slop Born from Human Slop?
If AI learned from us, then the genealogy matters.
Large language models (LLMs) were trained on vast amounts of human-generated content: articles, academic work, strategy documents, consulting artifacts, and certification content.
They did not invent the structural patterns within that material.
They absorbed them.
Confident generalization without sufficient grounding did not originate in silicon. Transferable abstraction existed long before generative models.
AI slop may simply be a compressed version of human slop – polished language and portable authority that professional environments already rewarded.
If so, the machine is not the origin story.
It is the mirror.
The Professionalization of Slop
Consulting Frameworks and Portable Authority
Consulting slop rarely looks careless.
It looks structured. It cites “best practice.” It references industry norms. But sometimes the person presenting it has never actually lived through the implementation of what they are recommending. The model sounds credible because it is familiar – not because it has survived friction.
Certification, Currency, and the Illusion of Experience
Certification can play a similar role.
A consultant promotes a newly updated credential. It signals currency. It signals alignment with the latest version of the framework. But the industry itself may still be wrestling with how that framework works in practice. The certificate confirms knowledge of the model – not evidence of applied experience.
In both cases, authority is reinforced by recognition.
Not necessarily by results.
If that kind of material fills our decks, courses, and documentation, it is no surprise that AI can reproduce it fluently.
When AI Slop Moves from Content to Operations
Content slop is irritating.
Operational slop is expensive.
Scaling Shallow Models into Systems
Consider a common example: an organization adopts a widely recognized incident prioritization model. It aligns with “best practice.” It is certified. It looks defensible.
But the organization has fragmented service ownership and inconsistent definitions of impact. The model is implemented anyway.
Within months, incidents are misclassified, automation routes incorrectly, and escalation paths conflict. The framework was not wrong.
It was insufficiently contextualized.
AI makes it easier to generate models quickly.
Scaling shallow models scales their weaknesses.
Slop at machine speed is manageable.
Slop embedded in systems is harder to unwind.
AI Slop: If Output Is Cheap, Discipline Must Be Expensive
If AI accelerates output, discipline must accelerate with it.
Before publishing or deploying something, we might ask:
- Where is this anchored?
- What assumptions are embedded?
- What constraints were considered?
- Who owns the consequences if it fails?
In an environment where output is cheap, reflection becomes scarce.
Measuring slop is not about shaming authors.
It is about protecting operations – and our own standards.
Closing Reflection on AI Slop
Before we blame AI for the rise of slop, we should ask a harder question.
Have we ever leaned on “best practice” to add weight to something we hadn’t fully tested?
Have we cited a certification to signal authority rather than demonstrate applied experience?
Can we truthfully say we have never used familiar language to make something sound more robust than it was?
If we are honest, most of us have.
AI didn’t invent slop.
It learned from us.
The question is whether we are prepared to stop contributing to it.
Ian Clayton
Ian Clayton is a veteran IT consultant, with over 35 years practical experience helping IT organizations transform their work practices and culture, from a technology focus to one centered on business and customer value.
Ian’s pragmatic approach enables professionals of any experience level address the tactical needs of day-to-day operations and opportunistic practice improvement, and the strategic thinking required by initiatives such as ‘digital transformation’.
As the author of the Guide to Universal Service Management Body of Knowledge (USMBOK™), Ian helps organizations and individuals understand where, and how to apply the original principles of service management as created by business product management. He helps establish better practices, and where required plot a personalized journey to a customer centered service provider organization.
He does this with the help of the same ‘outside-in’ thinking used by the most successful service organizations, focusing on a combination of service experience, workforce efficiency, successful customer outcomes, and levels of customer satisfaction.
Ian’s past practical experience spans the standup and operations management of Data Centers, product management of Sterling Software’s SOLVE: Automation product set, realignment of a worldwide customer service management operation supporting hundreds of thousands of end users and leading more than 50 implementations of the ServiceNow platform.
He is a recognized pioneer and thought leader in the service management industry, co-founder of the itSMF USA, recipient of the ITSM industry’s Lifetime Achievement Award for 2005, holds several important industry qualifications, and was selected to facilitate the first ITIL® V3 Expert classes outside of the United Kingdom. Ian also represented the USA in the ITIL V4 requirement sessions with Axelos product management under non-disclosure.
Ian lives in Sarasota Florida with his wife Delynn, and is currently employed as the senior advisor for ServiceNow platform solutions at Advance Solutions Corporation, like his previous employer Acorio, an Elite Partner. In his ‘spare time’, he is preparing the 2023 edition of the USMBOK, and updates to several practitioner guides including incident, problem, change, request, and configuration management.
