What can higher ed learn from AI-content generation?
One evening recently a thought popped into my head: What would happen if I gave creative control over a university brand and content project to AI-tools like ChatGPT and DALL·E 2?
Within an hour or two I had somehow become the unlikely founder of a new university:
What did I do?
Almost everything about the University of Anywhere has been decided, designed and created with the help of artificial intelligence.
I used a combination of AI-tools to help me create a homepage, study landing page, strategic vision (including video), research case study, and VC blog.
These AI tools delivered:
Visual identity - I asked ChatGPT what should be in the crest and it recommended an owl, a globe and a book. I fed this information into DALL·E 2 and it generated some logo concepts which I then quickly tidied up in Adobe Illustrator. I asked for Chat GPT for a university colour palette. It said "a classic and versatile colour that can be used for the University of Anywhere is blue" and "lighter blues suggest serenity and tranquillity, while navy suggests authority, intelligence and wisdom."
Copy - All the copy on the University of Anywhere has been written by ChatGPT. This includes the slogans for our UG campaign, the headline for our research promotion banner, the copy for every page and component, the VC blog and, well, every word you see on the site.
Images - All of the images on the site have been made using DALL·E 2 and, let's be honest, they are absolutely terrifying - but from a distance they are also eerily close to what you’d expect from a university website.
Video - The "Vision" video was put together using Lumen5 for AI-powered editing with a script written by ChatGPT and a voiceover via Murf.
The only thing that isn't really AI-generated is the site itself. This is something I quickly put together in HTML/CSS. My aim was simply to capture a basic "university" look that would accommodate the creative and content needs of my AI team.
What did I learn?
I've shown the site to the team at Pickle Jar and a few colleagues/friends from other universities and the general response is that it is terrifyingly accurate - and also just terrifying in its own right. This is especially true when you look too closely at the images.
There is a lot about the University of Anywhere website which does feel accurate to the status quo for university brands, websites and content.
It's tempting to focus on the "accuracy" aspect of this and be astounded or deeply troubled at what AI can do.
But I think it is far more interesting to focus on the words "status quo."
When I look at the University of Anywhere I feel like the AI has learned from the "as-is" of university content and communications. But that "as-is" is a world where almost everything created, produced and published is heavily diluted by internal politics, power struggles, strategic confusion, absence of trust and a lack of money or resource or time.
Teaching AI about university communications, content and marketing is like teaching a person what oranges taste like by only letting them drink a cordial that is 98% water and 2% orange flavouring. It's in the right sort of ballpark but it is nowhere near as strong as it should be.
I've seen quite a few headlines recently about "AI coming for your content/social team" and jokes on Twitter about replacing everyone with ChatGPT.
The irony here is that I think an AI-experiment like the University of Anywhere actually points in the opposite direction. The fact that AI can do a passable job of the "status quo" means that universities should be listening to and trusting their marketing, communications and content experts more than ever in order to move away from this tedious homogeneity and towards something far more strategic, distinctive and useful.
What next?
I've no idea what the future holds for the University of Anywhere - largely because I haven't asked ChatGPT yet. But in terms of the future of AI-tools, I think it's about understanding the roles they can and can't play.
I can totally see how they speed up prototyping and "getting something on the screen" when it comes to ideas development and first drafts. I can also see huge value in their research potential and the useful conversations that can be had with tools like ChatGPT. There's also going to be a huge range of very practical applications such as generating alt text for images or voiceovers for videos. It seems inevitable that as these AI-tools improve they will be a huge help for communications, content and marketing teams.
But as a strategist I'm also interested in the stuff that AI can't do. These tools aren't going to fix things like internal politics, power struggles, trust issues and strategic confusion - and these are the things that are often the biggest blockers to progress.
But perhaps, for now, they can help shine a light on why these things need resolving - otherwise everything just ends up looking like the University of Anywhere.
Need more?
If you need help with any of this, our team can help you. Connect with us for a chat.
