This week Barry Phillips considers how one employer is dealing with the problem of AI generated slop in the workplace.

Transcript:

Hello Humans!

And welcome to the weekly podcast that aims to review an important AI development relevant to the HR professional in around five minutes.

There’s a growing problem in workplaces, and it isn’t that people aren’t using AI.

It’s that some people are using it to produce industrial quantities of polished-looking rubbish.

Reports nobody asked for. Policies padded with paragraphs that sound authoritative but say absolutely nothing.

Welcome to the world of AI slop.

Now, let’s be clear. AI slop isn’t simply content produced with artificial intelligence. It’s content produced without sufficient thought, judgement or respect for the person who has to read it.

And if you work in HR, compliance or management, this matters.

Because a badly written policy doesn’t just waste time. It can create confusion, undermine trust and, in some cases, introduce risks that nobody spotted because everyone assumed the machine knew what it was doing.

So what should organisations do?

One employer, Clay, has introduced an AI writing policy built around four rather sensible principles.

Here’s the first.

Stand behind every idea and every sentence.

If your name is on a document, you own it. Not ChatGPT. Not Claude. Not whatever shiny new tool appeared on your screen this morning. You.

That means checking whether the content reflects what you actually think, whether the claims are accurate and whether you’d be comfortable defending every word.

AI can help you write. It cannot take responsibility for what you publish.

The second principle is that writing is thinking.

We often assume the purpose of writing is simply to produce a finished document.

But the process itself has value.

Skip that process entirely and you might end up with a polished document and absolutely no idea what it really means.

That’s not productivity. That’s borrowing confidence from a machine.

Principle three: respect the reader’s time.

Clay’s view is that you should spend more time creating a document than someone else spends consuming it.

Think about that.

If you spend 30 seconds typing a prompt, generate six pages and then expect five colleagues to read the whole thing, you haven’t saved time.

You’ve simply transferred the effort from yourself to everybody else.

And if your colleagues wanted an unfiltered answer from an AI tool, they could have asked it themselves.

The fourth principle is beautifully simple.

Longer is not better.

AI makes it ridiculously easy to generate more words. Unfortunately, more words don’t necessarily mean more insight.

If the original prompt says everything your colleagues need to know, consider sharing that instead.

Now, the important point here is that Clay isn’t banning AI.

It isn’t shaming employees for using it.

It’s setting expectations about judgement, accountability and consideration for others.

And that’s exactly where this conversation needs to go.

The real test of AI in the workplace isn’t how quickly we can fill a page. It’s whether what we produce is accurate, useful, genuinely understood and worth someone else’s attention. 

Use AI by all means. But own the thinking, do the checking, cut the waffle and respect the reader. 

Because if artificial intelligence makes us faster at producing nonsense, we haven’t become more productive. 

We’ve simply found a more efficient way to waste everybody’s time.

That’s all for this week. Until next time bye for now!