Want a healthier workplace? Don’t fear AI, fix your feedback systems

As people use AI more, employers are being overwhelmed by a growing tidal wave of complaints, cover letters, e-mails and more.

For example, a new Princeton University study quoted by The Economist shows that the average cover letter length has doubled since ChatGPT came online.

And using AI for this stuff can be for the better – it can be useful for taking sincere, but unpolished thoughts and articulating them in a way that makes sense to other people – especially important for those who otherwise find it hard to articulate themselves.

But it can also generate reams of nonsese: high word count, low meaning slop.

The temptation is to clamp down on the use of AI – I recently saw a suggestion that employers add additional steps to their workplace complaints process, if they think an employee used AI to prepare their complaint.

I understand where the idea came from – it was trying to cope with a growing volume of nonsensical complaints that create lots of work for the people receiving them.

But the real problem here is nonsensical writing, not AI.

Let’s take the example of complaints.

Imagine the government gave everyone unlimited free and high quality legal advice whenever they needed help submitting a complaint. Such that any employee who had a legitimate issue could submit a sensible and articulate complaint any time they needed to.

Because the friction involved in submitting a complaint would go down, the volume of complaints would surely go up.

And this would create real short-term challenges for employers as they adjust to the increasing volume of complaints.

But it would be entirely inappropriate to put in place additional steps, just because someone had received help from a lawyer.

It would be wrong – we shouldn’t be designing policies and systems that discourage legitimate complaints, just to make our lives easier. And it would disproportionately affect the most vulnerable staff who have the least power, confidence and capability to assertively articulate themselves – these are the people who might benefit the most from reduced friction (e.g. through lawyers, unions, AI etc.).

And it would be dumb – sensible and legitimate complaints are a gift to employers who are serious about creating engaging and impactful workplaces.

Does that mean we just have to grin and bear an explosion of AI rubbish? No. But we should focus less on the tool (i.e. AI) and more on the actual problem – nonsense.

And there’s an upside here. If AI can help people present their grievances more clearly – grievances that exist whether or not people feel confident complaining about them, organisations will benefit from detecting them earlier.

Because more complaints doesn’t necessarily mean a worse culture. It might just mean that people are finally able to articulate issues that they previously felt powerless to speak up about.