What do we owe our audiences (and ourselves) when AI is involved?

For those of us in marketing and communications, the question of whether to use AI is no longer much of a debate. The technology is here. People are using it, organizations are experimenting with it. And so the conversation is moving from “Should we use AI?” to “How should we use it?”

This question has been thrown into focus by the recent news that, thanks to EU regulators, Claude will start watermarking AI-generated text. My initial gut reaction is one of cautious enthusiasm. I’m not particularly bothered about catching people using AI, but I like the idea of making its use more visible.

This is an issue worth thinking about, particularly for anyone responsible for communicating on behalf of an organization. We spend a lot of time talking about authenticity in communications. We tell brands that their success hinges on trust. We encourage orgs to show their work, admit mistakes and communicate like people rather than faceless institutions. So it would seem more than a little disingenuous to treat AI as something that should remain invisible.

I don’t think using AI automatically makes communication less authentic, nor do I think every AI-assisted email, social post or piece of copy needs a disclaimer. There’s a difference between using AI as a tool and asking AI to do our thinking.

But the reality is that audiences have a legitimate interest in understanding how content is produced. Which means that organizations have an interest in establishing some clear norms before they find themselves explaining their AI use defensively. That’s where I think the watermarking could be useful as a nudge toward greater openness. If AI-generated content is more easily identifiable, it becomes harder to pretend that the distinction between human and machine-produced work doesn’t matter.

There may also be a useful side effect: a little friction. If we know that AI involvement can be detected, perhaps we’ll think twice before using it for something we could easily do ourselves. Perhaps we’ll ask whether AI is actually adding value, or just making the process faster and more convenient. 

My personal goal with AI use, and my guidance when advising clients, is to use it sparingly, with intention – and to be honest about where it fits into the work. Because if trust is what we’re trying to build with our audiences, transparency matters. So does trust in our own judgment. We’ve spent our lives and careers developing the ability to think critically, make decisions and communicate with purpose. These are skills worth protecting rather than automating away.