This blog is not written by AI (but the summary is). And why this matters in engagement.

This blog is not written by AI. Can you tell? Do you care?

As someone who prides themselves on their long-honed and much practiced writing skills, I have a somewhat love-hate relationship with AI.

I don’t want it to do the writing for me, because frankly, I think my writing is better (poor grammar included), but if it helps me to organise of a jumble of incoherent thoughts and makes the final edit a little easier, well, that is ok by me.

I see a lot of comments about “AI slop” and, if we are being honest, there is a good dollop of writer snobbishness in this. Which I understand. Can I spot AI written content? Yes, most of the time. Does it bother me? On the whole not really, certainly not compared with the human-written nonsense we sometimes read. In fact, as someone whose job it has been to turn rambling incoherence into smart narrative, sometimes I am grateful that AI has helped them to understand what it is they are trying to say in the first place.

However, this does not mean that getting AI to generate all your content for you is ok.

Note that word “generate”. I have chosen it carefully, advisedly, and with thought and attention. AI does not write. Not in the way a human writes. It generates copy based on a set of rules coded into it. Writing requires consideration, care and emotion. AI has none of these skills.

It doesn’t even have grammar perfectly down – have you ever spotted the random capitalisation in AI work? But then neither do I… (see what I did there with that opening “But”…? – one for the writers in the room perhaps…).

If I make grammatical errors, there are two things going on: firstly, like that “but” and the “and” in the blog title, I am deliberately breaking standard rules to make a point, create a flow, etc; secondly, and most likely, I just made a mistake because I am human and the last time that I remember being taught grammar I was 15 and my English teacher was patiently and mostly unsuccessfully explaining to a room of GCSE students how to use the possessive apostrophe.

AI can do the first of these things – especially once it learns your writing style, it can create new rules for itself that override the standard ones. However, the second it can’t do; it cannot make mistakes because it is human.

It is not human. It does not know how to be human. It simply knows how to generate copy. And asking AI to make copy sound more human can never make it feel more human. It is still computer generated.

And this is the vital point. As most communications professionals know and is really at the route of their AI despair.

The human touch

People choose people. People buy from people. People work hard because they care about the people they work for or with. People respond emotionally to and are persuaded by other people. I can come up with another 10 ways of saying this, but I am sure you get the point!

It is the “people” in it that counts. In other (more grammatically correct) words, it is the human touch that makes the difference.

As engagement professionals, we aren’t there to persuade people of things. But we are there to create emotional connections and trusts. We are there to make people feel safe to share their deeply personal experiences with us, to convince them to give up their time to consider and respond to ideas and options that are being put forward. We need them to believe that we care about what they tell us and that we will share their feedback faithfully and without manipulation.

When we are trying to create an emotional connection, some studies show that people prefer human written content – even if they actually find it easier to read the AI written content.

So having the human touch in a narrative is important. Let’s be honest, we all know that we have switched off halfway through a dull document that feels copy and pasted from something else and devoid of all emotion. And that is when a real person has written it. How much more likely are we to do that if no actual human was involved?

AI disclaimer

There is another point beyond who (or what) wrote it to consider though – and perhaps it is the more important one. That is the declaration of who wrote it. Are we nearing the point where we need to add content production disclaimers and handling intent disclaimers onto documents?

For engagement professionals, this is hugely important, as honesty and transparency are essential to building and maintaining trust. It particularly relevant in the engagement and consultation reports.

If you know you are going to use AI to help you code, theme and sort through 1,000+ responses, should you declare that upfront when you are doing your engagement? Perhaps in that handy “next steps” section that appears in all good engagement work and shows the intent to actually do something with the information collected?

Is it enough to merely declare it after the fact in the report? Something like, “AI was used to support the review of the feedback, but all responses were read by a human and everything was manually checked and the report written by the team.”

Or do you just use it and forget about it?

Most of the blogs and content that are AI generated generally take the latter approach, but in the engagement world, I think this is a dangerous tactic to take. If it comes out later that AI was used, it could look disingenuous, and it could lead to some interesting questions in court (should it be challenge) about how conscientious consideration can be proven. The Consultation Institute has an interesting blog on this element, if you want to go into more detail on this.

I would argue that the absolute minimum is a statement in the report of if and how AI was used – including if it wasn’t. Increasingly, with the AI pendulum swinging further, I would advise that if you know you are going to use AI in any part, declare it up front in the engagement information. In the opaque and governance-ridden world of public sector decision-making, cover-ups and lack of transparency are easy arguments to make, so why gift an opportunity for challenge and disgruntlement. Being honest also helps to build and maintain the trust that is vital to the work.

The easiest solution – and the one I would recommend – is that you publish a section on your engagement pages that sets out how you use AI in engagement. This gives you space to be quite specific and detailed, rather than trying to cram it all into 2 sentences within another document. You can then point to this in all your activities and demonstrates a commitment to openness.

Is this something you would consider? Where is your organisation and team in using AI and what do you declare?

Open Voice Lab recommendation

At OVL, we can help you look your project and decide the best way to use AI to deliver meaningful and effective engagement, show you some techniques, and help you draft an overview for your website.

Reach out to have a conversation with us.