The Rise of AI Workslop: Is AI reshaping the workplace for better, or for worse?

A friend recently told me about her experience trying to hire, and how candidates had used AI to create what looked like promising presentations. However, when she interviewed them, they could not elaborate on the content.

Colleagues, get ready – “workslop” has entered the chat. This term, coined by researchers in the Harvard Business Review, refers to low-quality, AI-generated work that looks and sounds OK at first glance but, upon closer inspection, doesn’t actually mean anything. Worst of all, it usually results in someone else – possibly you – having to do more work to clean it up.

“Young people are really embracing [AI] and bringing it into the workplace without necessarily understanding all the risks involved,” says Simon Kemp, founder of Kepios, a Singapore-based advisory company that analyses what people really do online. “And I think management hasn’t taken the time to understand the implications properly.”

As a result, workslop (and AI in general) is showing up in all sorts of ways, from presentations to emails, which Simon finds particularly aggravating. “You have people using AI to write emails, then people receiving them and using AI to summarise them, and then using AI again to reply; in the end, it’s just busy work,” he says. AI is allowing people to get away with busy work on a much larger scale, which has become a problem. It can also lead to confidentiality issues if people are not careful.

AI as a tool, and not a replacement for actual thinking

Simon likens working with AI to meeting someone in a bar. “If they tell you something, you wouldn’t just take it at face value and go back into the office the next morning and use it,” he says. “So I think we’ve got to treat AI a little bit like the dude we meet at the bar.”

Ultimately, he stresses the importance of remembering that AI is a tool, not a replacement for actual thinking. “Within teams, that nonsense is something we have to deal with. But if you’re working with a client or customer, or communicating with the government and tax authorities, you’re exposing yourself to incredible risk if you’re not reading things properly and then formulating a correct answer that’s actually been checked by your brain.”

“AI always tries to please,” adds Bhavish Advani, Of Counsel in Watson Farley & Williams’ Dispute Resolution Group, who handles cross-border commercial disputes. He also works as part of the law firm’s Innovation Network, which develops guidelines for technology adoption across its global offices. “AI tries to find ways to say yes. It rarely takes a definitive position, even if it can weigh pros and cons — that’s why in many policy debates, the focus is on ensuring it doesn’t become a decider or determiner.” He emphasises the importance of knowing your source material, and if something sounds too good to be true, it probably is — a clear red flag.

If you’re going to use AI in your work, he advises explicitly setting limitations in prompts and being especially cautious with confidential information. “Even if you’re only using AI tools for referencing, formatting, or proofreading, this should be done without any information making its way into the public domain. Enterprise versions offer some assurances — privacy controls are never available with free versions. Additionally, final content should always be cross-checked,” he says.

How to use AI the right way

Chris McDaid, Director of Retail Data Science at The Tecsa Group, a consulting company focused on loyalty programmes, and his team use AI tools regularly to create code, build reports, and brainstorm. He has seen the benefits of using AI for idea generation and coding.

“We’ve found it works best to treat AI like a junior analyst with great ideas,” he says. “It can come up with some really useful stuff, but it still needs a lot of guidance. It’s never going to be right all the time.” He also cautions about mistakes and inaccuracies. “You always have to sense-check and make sure it’s telling the truth,” he adds, describing a time when it wrote code but didn’t accurately describe what the code could do.

Simon agrees that there is a right way to use AI at work and believes it will only improve in the future. But for now, we are still in the “wild wild west” of understanding its usefulness. “[Large Language Models] have only been accessible to the general population for three years now, so we’re still just getting our heads around it.”

He adds that, based on recent data, roughly three-quarters of AI use is still for personal reasons (such as seeking companionship!). “So only a quarter of its usage is in the workplace.” He believes it will take time to see where it all fits together, and in the meantime, there will be more short-term horror stories. “But I certainly see more benefits in the long run, partly because all of the negative bits will become horror stories, and we will realise that this is not the way to use it.”

Should we worry about our jobs?

We’ve all heard the stories that AI is coming for our jobs, but at the end of the day, it’s important to remember that it’s a tool, not a solution. “AI certainly automates certain tasks. But that doesn’t make a job redundant,” says Simon. “There are circumstances where that will be the case — and I think we need to be very careful about how we manage that as societies and as industries.” He gives the example of the laptop, which essentially eliminated the role of typists. “The people in those roles had to evolve and apply their skills in different ways. So it’s also about understanding your value and how that can help you move to the next level.”

And while there’s a lot of talk about how AI is revolutionising the workplace, and how businesses will be left behind if they don’t use it, Simon points out that everyone using AI is basically using the same tools. “So really, it’s not a competitive advantage if everyone has the same thing.”

Remembering that AI is a tool — not the solution — is a good way to ensure you use it properly. “The advantages and disadvantages of it have much more to do with how we make use of it,” says Simon. “AI can be incredibly valuable in the same way that laptops and mobile phones are if they’re used appropriately. But buying a saw doesn’t make you a carpenter. Similarly, having access to AI doesn’t make you any of the things the AI might be able to do until you develop that skill. Until then, it’s just a fancy tool in your toolkit.”

In part two of this two-part feature on AI workslop, we share tips on how to spot AI intervention at work... from AI itself. Click here to find out more.