We wrote an AI policy. Here's what's in it | BlueSky PR
3:25

Who wrote the last piece of content you read? Was it a person? Did you skim through it? Or really focus? There's a decent chance it was written by no-one. No person anyway. But instead by an AI model trained on the average of the internet that produced something plausible, someone pasted it in, and it went live. It read fine. It also said nothing, sounded like everyone else, and was forgotten by lunchtime. Because who wants to invest time reading fine? I for one want interesting.

We think the question of who writes your content matters more now than it ever has. So we write down what we believe, and made it public. Here's what's in our AI policy, and the thinking behind it.

What do we actually believe?

Our position is simple. AI is a useful tool for distribution, research and efficiency. It's a poor substitute for someone who knows your market. The voice behind a recruitment firm's content should understand your clients, your candidates and the pressures your sector is under this quarter, not predict the next likely word. Generative AI is very good at sounding like an expert. But real experts have passion, opinions and something you want to read.

When anyone can generate a thousand words in seconds, the value of generated words is zero. What cuts through the drivel are specific, honest and useful pieces of content, and those are still made by people like you and me.

Why write a policy at all?

PR runs on trust. Journalists trust that what we send them is true and ours. Clients trust that their embargoed news, their media lists and their unpublished results are safe. Audiences trust that a byline means a person stood behind the words. Used carelessly, AI can easily erode all three. A policy is how we make sure it doesn't.

What's in it, in plain English?

The full policy is on our site. It commits us to a few things that we think are pretty important...

A named person is accountable for everything whether AI has touched it or not, and nothing goes out without human review.

If AI has shaped substance or style, we say so. Where it's checked a spelling, or suggested some SEO improvements, we won't pretend otherwise, but it’s less important to highlight.

Every fact an AI suggests is checked by a person before it's published, because confident and wrong is the LLM's natural state.  And we like to be confidently right.

We protect confidential client information, media lists, journalist contacts and embargoed material.

We don't use AI to imitate a journalist's style, mimic a named author, or claim authorship and expertise that isn't there.

And we're clear about where we feel that AI genuinely helps: ideas, outlines, desk research, competitor summaries, and sentiment or engagement analysis. Tools, in other words, not authors.

AI is an aid, not a substitute for professional editorial, strategic or creative judgement. 

Why should you care?

If you run a recruitment firm you'll need a version of this yourself before too long. Your staff already use these tools daily, clients will start asking, and "we hadn't thought about it" is a weak answer when something goes wrong. We've turned ours into a template you can adapt, so you're not starting from a blank page. You can download it here.

 

We'll keep ours under review as the tools change. But the principle underneath it won't. Good content still comes from people who know what they're talking about. AI just helps more of the right people hear them.


 

Author: Jennifer Wright

As BlueSky PR’s Marketing Director, with almost 10 years’ prior experience in the recruitment industry, Jennifer writes articles and guest posts to inspire recruitment agencies to build their brands, improve their content, bring in more leads and generally make their lives easier.

 

 

Share This Post

f X in P @ W

Post Your Comment