When imperfection becomes your competitive edge

It has always been human nature to try and improve things. As close to perfect as possible. Assuming ‘perfect’ involves some degree of subjectivity in most areas (especially the creative sector) we might at least agree that perfect means, ‘meets the brief without error’. 

In many fields, this has been possible:

• Well-trained, experienced photographers taking well-composed photographs with good technical execution (in focus etc.)

• Interesting and useful articles with a cogent structure, good signposting and clear language

• Illustrations that show stylistic and artistic flare while clearly and cleverly bringing an idea to life 

But, with the proliferation of AI-generated content - from text to image and video - will imperfection become a strength?

TL;DR: We once strived for perfection. Achieving it was rare. Now we have generative AI, so everything is perfect - and with little effort. Will imperfection become a point of difference?

The road to perfection

People already complained about how stock photography was dull, generic and, subsequently, made businesses and their content seem generic. However, stock libraries allowed businesses of all sizes to access perfect photos of perfect people in perfect offices having meetings. 

Until Nov 2022 at least the content that businesses created would have been written by a person. Now, ChatGPT has allowed businesses to create ‘perfect’ articles and product pages - apt for their business, without error and for FREE. People who said ‘now everyone’s using GPT to write content, it’ll mean the web is full of crap content’, but the web was already full of crap content - and that was all written by humans. However, it does mean that many businesses will have text that’s equally generic to match their generic images.

And the allure of the quick, easy and free is SO alluring that many businesses will find it hard to resist - especially in these challenging economic times.

For more than a decade, I’ve been sharing this graphic with colleagues:

This Truth is no longer true. The ‘impossible utopia’ has arrived. At worst, because Great is a value judgement, so you might say that we are in a new intersection of Fast and Cheap: instead of ‘dipped in ugly sauce’ it might read ‘generic’. But it is hard to deny that it is possible to create great text and images with AI (and, increasingly, video). In fact, you might argue that the small grey ‘free’ circle should be where Cheap is.

The near future

So, a few brands will have the budget and appetite to continue creating unique, human-created images and text. The big brands like Nike and Burger King who are completely like you or your clients. But many will turn to our robot overlords for their help. This will result in a lack of difference between brands - everything perfectly written and perfectly generic. All loaded onto a templated website designed for businesses in your sector.

But is there a sweet spot in the middle?

Competitive advantage

But we don’t have to be Apple to be unique. And maybe, in a world or machine-made content - mistakes become a competitive advantage. The thing that catches the eye, the wink and the nudge, the shibboleth that shows the reader that this is actually written by a person.

My instinct was always to edit ‘um’ and ‘ur’ out of podcasts. But I read that these ticks actually help to refocus the listener because they’re unexpected.

In 2003 electronic music producer Matthew Herbert published his Manifesto Of Mistakes. Dance music has been accused of many things: easy to make, lacking in humanity. However, it is true of all music that it had been becoming more and more sanitised: removing mistakes and correcting errors. Auto-tune, easy sound design software and computer editing made it easy to make everything sound perfect. But, many of us enjoy the sound of 4 people in a room playing instruments, rather than the product of a hundred takes and a thousand edits. 

And it wouldn’t be a blog from me if I didn’t mention writer William Burroughs, whose prose frequently included spelling errors and lacked punctuation.

Many would argue that to be perfect, content must contain the human element - whether that’s a unique opinion, idea, emotion or a mistake.

Will the typo become the sign of human content? Will we be reassured by small mistakes? And, as such, will writers insert errors into AI-generative content to make it feel more human?

Maybe. 

How will we tell? AI may well get better at mimicking humanity and even develop the kinds of AI personalities that science-fiction has promised for nearly 100 years. But we’re not there yet.

Authenticity

The guiding principle should remain authenticity. The key will be to train your robots to deliver content that feels authentically you - like you did when you commissioned writers. 

I wouldn’t presume to tell you how to create your content - whether that’s text or image - as long as you promise to pass it through the usual quality checks: do the ideas feel fresh and unique? Does it look/sound like someone who works here created it? Would I have happily paid someone to create this five years ago?

Don’t compromise because AI is ‘cheap’. Train them up. Push the envelope. Demand more from your robots.

Remember, a bargain isn’t a bargain if you don’t actually want it.