Why letting AI write your Arts Council England Application (unsupervised) is a terrible idea.

Look, we get it. You're staring at a blank Arts Council England application form and the thought of asking ChatGPT or Gemini to just... write the thing for you is incredibly tempting. AI can knock out polished-looking text in seconds. Job done, right?

Not quite. Assessors can spot AI-generated applications a mile off, and if your application is so generic they can't tell what you're actually doing, you're in trouble.

We're going to show you why unsupervised AI writing is risky business, using Pink Stitch as our case study. Pink Stitch is a creative health project that was successfully funded by Arts Council England without any AI involvement. For this blog, our Junior Consultant Janeen Loh ran an experiment: she took the successful application and put it through both ChatGPT and Gemini, asking them to write sections from scratch and review what was there. The results show exactly what AI gets wrong and why relying on it can seriously damage your chances.


The Problem with AI

AI doesn't actually understand your project. It's predicting the next most likely word based on patterns in massive datasets. So whilst it might sound fluent and professional, it's not grasping the nuance of what you're trying to say or whether what it's saying is even true.

ACE have cottoned on to this. In their 2025 position statement, they pointed out that AI can lead to submissions with "similar language and text" that "distract from an application's unique proposition." If your bid reads like it came from a template factory, assessors will smell a rat.

ACE want to know that your proposal is "true, original and deliverable." Funders value authenticity over polish every single time.


Common AI Cock-Ups That'll Sink Your Application

Based on the Fundraising Everywhere presentation on "AI doozies," ACE's guidance, and our Pink Stitch experiment, here's what AI gets wrong:

  • AI loves phrases like "our innovative project will engage the community and transform lives", which says absolutely nothing specific. When we asked AI to review Pink Stitch, both ChatGPT and Gemini flagged this immediately. If your application could describe literally any project, you've got a problem.

  • Transformative. Cutting-edge. Impactful. AI absolutely loves dropping in flashy terms without context. When we asked AI to suggest improvements to Pink Stitch's narrative, Gemini kept recommending phrases like "safe space" for breast cancer survivors without explaining what made the creative approach actually special. A human writing authentically would naturally include those personal insights rather than leaning on tired buzzwords.

  • AI doesn't know your specific partners, your particular community, or why your approach is unique, so it defaults to the vague. When we asked AI to write an evaluation plan for Pink Stitch from scratch, the result was worryingly generic, lots about collecting feedback and counting participants but nothing about artistic quality or creative outcomes.

    ChatGPT's review of the existing evaluation immediately flagged this, noting ACE would expect to see how you'll measure artistic success: participant creative output, audience response to artistic merit, reflections from curators, not just health statistics.

  • When we asked AI to draft sections of Pink Stitch emphasising the health angle, the text got way too bogged down in cancer statistics whilst barely touching on the artistic vision. ChatGPT's review spotted the problem immediately: "ACE occasionally rejects creative health projects if assessors feel the 'artistic idea' is overshadowed by health messaging." A human writing the application would instinctively balance these elements.

  • Pink Stitch's venue hire was £13,680. When we asked AI to write budget justification, it listed the cost but didn't explain it properly. Both ChatGPT and Gemini's reviews explicitly warned that such a large venue cost would alarm assessors without proper justification. The fix was simple: explain what that fee includes (equipment, technicians, security) and why that specific venue is critical. AI won't know to do this unless you tell it to.

  • When we asked AI to draft project management sections for Pink Stitch, it didn't highlight safeguarding measures at all, despite the project dealing with health and trauma. ChatGPT's review immediately caught this gap, noting that "ACE increasingly expects safeguarding considerations in creative health work."

    Environmental responsibility is another area AI skips. These might seem like minor details, but to ACE they demonstrate professionalism.

  • ACE's position statement warns about "systems that retain users' information for training commercial models." If your proposal contains sensitive details like personal stories from participants, bunging that into a public AI tool could violate privacy or copyright.


Why Your Authentic Voice Wins Every Time

Your authentic voice and creative vision are your application's greatest assets. If you let AI write without firm guidance, you risk submitting something that sounds robotic, impersonal, or eerily similar to dozens of other applications.

Pink Stitch succeeded because it was written by humans who understood ACE, knew the sector, and could craft an authentic, compelling narrative. The project's unique voice shone through.

So Should You Never Use AI?

Not necessarily, but you need to be careful. AI can be useful for getting past writer's block or as a rough first draft. The key is supervision and editing. Think of AI as a keen but inexperienced assistant: it can generate raw material, but you need to refine it, inject the personal touch, and double-check absolutely everything.

Before you submit, make sure the language sounds like you. The narrative should be rich with your own examples, and all the critical details should be clearly there. It's far better to submit something slightly rough around the edges but genuine than a slick essay that rings false.

Remember: funders value authenticity over polish. A human assessor will forgive the odd clunky sentence if the passion and originality are evident. They won't forgive a soulless, generic proposal, even if it's stuffed with the right buzzwords.


Coming Soon: Our ACE Application Review Service

We're excited to announce that Alira Arts will soon be launching a brand-new ACE Application Review Service, created specifically for artists and charities applying to Arts Council England. We'll be opening with a limited number of places, and we expect them to fill quickly.

If you want professional feedback on your draft, guidance on tone, structure, or ACE-specific expectations - don’t wait.


Final Thoughts

Unsupervised AI writing can introduce generic prose, skip critical details, and raise all sorts of red flags in your Arts Council application. Our AI experiment with Pink Stitch revealed just how easily an AI-drafted bid can go astray: muddled artistic focus, unexplained budget lines, cliché-ridden text, the works.

By approaching AI tools with caution and using them to assist rather than dominate, you can protect your application's integrity.

In our next blog post, "Using AI the Right Way: Strengthening Your ACE or DYCP Application," we'll flip the script and explore how AI can actually be helpful when used properly, as an editor and reviewer, never as a wholesale ghostwriter. Stay tuned for practical tips on making AI your ally, and why expert human fundraising support is what actually gets projects funded.

 

Keep Reading

Previous
Previous

Using AI the Right Way: Strengthening Your ACE or DYCP Application

Next
Next

Using the Indices of Deprivation as a Planning Tool in the Creative Industries