Crowned Macquarie Dictionary’s Word of the Year for 2025, ‘AI slop’ isn’t just a buzzword, it’s a problem. For those unfamiliar, or fortunate enough to not have experienced it, AI slop is cheap, mostly-automated AI content that fills space and grabs attention but offers little substance, accuracy or original thought.
Research cited by Harvard Business Review shows that 40 percent of full-time employees across industries in the United States report having received AI slop, occurring mostly between peers (40%), sent to managers by direct reports (18%) and flowing down from managers and more senior staff to their teams (16%). For an organization of 10,000 workers, given the estimated prevalence of slop (41%), this equates to over US$9 million per year in lost productivity.
It’s permeating social media too, with AI generated video content flooding video sharing platforms. It is also increasingly being used in LinkedIn posts, with analysis by Originality AI revealing that more than 54 percent of longer English-language posts on the workplace social media platform are likely AI-generated.
Such is the threat of AI slop to productivity, a market is being carved out for companies willing to solve it.
In October last year, Australian startup Enhance Labs, a voice-first collaboration platform that amplifies human thinking over AI, raised US$1.6 million, with companies including Amazon, Canva and Microsoft named early adopters.
Debating where the fault lies
The obvious place to start is deciding it’s the fault of the technology or those with their hands on the wheel. I am a firm believer that the real risk of AI slop isn’t the use of AI or the technology itself, but rather the absence of clear human-in-the-loop accountability once AI enters the workflow. That accountability is part ownership and part skill set – which data attests.
Skillsoft’s 2025 Global Skills Intelligence Survey shows critical skills gaps exist, with 91 percent of HR professionals believing employees overstate their skill proficiency, particularly in leadership AI, and technical domains; 41 percent say resistance to change is the top barrier to AI adoption; and 28 percent cite a lack of technical AI expertise. People can’t be expected to be accountable for AI and its outputs if they don’t have the skills and understanding of how the technology operates.
I am a firm believer that the real risk of AI slop isn’t the use of AI or the technology itself, but rather the absence of clear human-in-the-loop accountability.
Addressing these gaps means leaders and managers will be better equipped to understand the nuances and flaws of AI – including spotting when AI slop is creeping in. It will also lift the floor by training the entire workforce to utilize the technology, meaning a better standard of output from top to bottom, from creation to review to actuation.
It’s also important to treat AI like you would an intern or junior employee. Human employees make mistakes, do not always follow tasks with complete accuracy – often the fault of the brief or prompt – and won’t always get the tone or copy right when developing materials.
AI agents are exactly the same. Treating them as such means employees do not default to thinking the outputs of the technology are correct without thorough oversight.
Developing an antidote
Developing a skilled workforce – a ‘skillforce’ – is a potent antidote to AI slop and requires multiple layers. Perhaps most critical is understanding how the talent and skills landscape has shifted from a job-centered architecture to a skills-based one. With the proliferation of AI – which is capable of picking up menial, operational driven tasks – the value of more specific cognitive skills has never been higher.
With the proliferation of AI – which is capable of picking up menial, operational driven tasks – the value of more specific cognitive skills has never been higher
Recognizing this can be incredibly transformative, not just to business, team and individual performance, but also the overall output and quality of work from both human and AI talent. It transforms how you select skills, how you measure them, and how you deploy them.
At Skillsoft, for example, we’ve gone through our leadership team and pinpointed five skills they need to be a successful leader in the company. This provides a framework for performance and measurement and makes performance reviews and succession planning much simpler.
We’ve now included this framework in our hiring process, creating a whole new interview guide for any incoming executive outlining the five skills they should have.
Start small to go big
To adopt this model, it’s important to start small, integrate it everywhere you can, understand it, learn it and pilot it. Starting with the leadership team means you can effectively trial it over a smaller cohort before rolling it out to every level of the business, where skills and work context may vary depending on seniority level or department.
It’s also vital to constantly evolve and iterate. For example, three of our five core leadership skills may be redundant this time next year. So, like any effective strategy or process, constant reviews and fine-tuning based on where you are in the moment is key.
An important thing the rise of AI slop has done is highlight the power of human-centered skills like critical thinking, analysis and empathy – and how it must be nurtured at every level of human talent. Empowering human talent to focus on these skills can not only address poor AI generated work in the immediate term, but also protect the future of your business and its people.
When treated with thoughtlessness and without the necessary care, AI can be more of a hindrance than a help.
According to the World Economic Forum’s ‘Future of Jobs Report 2025’, soft skills – otherwise known as ‘power skills’ – which AI agents can’t learn, will increasingly rise to the top. Harnessing the human-centric and interpersonal skills that AI lacks is the best tool we have when it comes to generating rewarding, harmonious work.
They say you can’t fight fire with fire, but there’s also real value in using AI to combat AI sloppiness, with tools now available that make it possible for organizations to publish interactive, practice-based learning experiences in minutes rather than weeks.
Traditionally, this process has been slow and resource-intensive – but now it’s possible to move from design to published, enterprise-ready learning experiences in as little as 15 minutes, compressing timelines that once stretched on for months. This makes fast, personalized, skills-centric learning a reality.
End the slop
There are significant and valid doubts about ROI on AI tools and the quality of the actions and materials the technology generates. So much so that phrases like ‘AI slop’ are commonplace in the everyday vernacular.
When treated with thoughtlessness and without the necessary care, AI can be more of a hindrance than a help. It can mean staff having to redo work, which can mean the time it takes to complete a task or project can take longer with AI than without it. Or worse, it can dilute and impede customer-facing innovation and thought-leadership eroding trust with customers, key stakeholders and the wider public.
This is where skills management comes into play. Building a skillforce that recognizes the human-AI centered world we now work in, and the evolution of legacy job titles to future-facing skills, is fundamental to addressing AI’s pitfalls and utilizing the technology to its full potential.
Doing so requires starting small – like a leadership team or specific department – before wider rollout. Start by incubating the skills you need in your business to the most granular detail, embed thorough measurement processes to track those skills over time, and then everything you do from hiring to nurturing and upskilling talent will be framed through that lens.
The result? Better work, engaged employees, improved productivity and less slop. I call that a win–win.
