Let Machines do the Machine work
I’ve been noticing a flood of social media posts, banners, and graphics from churches across Manchester and the wider Church of England lately, and the telltale signs of Canva’s AI generation are everywhere. Not only do these images suffer from a chronic case of over-engineered mundaneness, but they all look exactly the same.
I completely get it. The social media landscape is relentless, and the temptation to quickly scramble a post together and throw it out there is real. I see a solid use case for these tools, and I’m just as guilty of needing a quick fix when the pressure is on to share good content.
But leaning entirely on generation creates a twofold problem:
First, AI “slop” is so easy to identify now that it is actively hurting reach. Research suggests it is far less effective at engaging audiences, with social platforms now actively labelling and deranking content if AI is involved.
Second, it strips out the unique human touch, which is sometimes the only thing a church or brand has to genuinely connect with its community. It is a lot like the current state of fast fashion. When every major brand regurgitates the exact same style, everything blends into a beige blur. The ones that actually stand out are the ones that decide to rebel and just be themselves.
To borrow a line from Syndrome in The Incredibles: when everyone’s super, no one is. We now have the power to generate endless content while sipping a cup of Earl Grey in our pyjamas before we’ve even got out of bed. The catch is that the output is completely sterile.
People are starting to try and mask their AI usage, mostly by desperately avoiding obvious tells. We are seeing the digital extinction of the em dash, which is now heavily associated with ChatGPT output regardless of who actually wrote it (there was even a case in the US where a professor accused a PhD student of using AI purely based on an em dash). This paranoia has also led to an exhausting overuse of the comma, where people are breaking up sentences unnecessarily just to prove a human is at the keyboard.
So, what is the fix?
We need to leave machine tasks to the machines and stop trading human taste for convenience. Let’s work smarter. I am not saying we abandon these tools. But what would it look like if we actually fine-tuned the results to introduce genuine personality, theology, and specific parish flavour?
Try using GenAI to get you 60% of the way there. Let it handle the framework, the concepting, and the heavy lifting. But that final 40% (the actual voice, the brand, the final polish) needs to be placed by a human hand.