Human v machine: who will win our competition between an AI and a real writer?

Personally, I’ve spent hours prompting DALLE-2 and sitting agog while Lex finished my sentences. The launch of ChatGPT, has brought the tech to a wider audience, with interest levels going through the roof.

At Nucco, our discussion has already moved on from whether it’s cool (it is), to 'how good or useful are the tools in reality?'

So we decided to set a test. Human v AI. Human and machine asked to complete the same task: the writing of a short article examining “The power of AI in creative industries”.

Thanks to The Drum, we’re going to be publishing both of those posts in the coming days and using views, social shares and likes to gauge which is the most valued. Keep an eye out for the follow-up article about who performed best, but for now let's do some pre-scoring to see who's in the lead ahead of the competition.

What can you expect from AI tools today?

AI cannot yet imagine for itself. For example, it didn’t come up with the idea for our creative competition. A human had to do that. And the software then needed concise prompts and some level of pre-thought before it could start. So in many ways that’s 1-0 to the meat bags in the very first minute.

That said, the launch of ChatGPT has made prompt writing much easier. If you’re unfamiliar with the software, it’s most easily explained as using one very exciting AI tool to write the instructions for another.

Beyond imagination, we've also noted a difference in insight and practical examples between human and AI writing. AI’s seem to struggle to reference both other tools and other authors.

2-0 to the flesh sacks.

The machines aren’t done

From this point, the competition gets a bit harder. That’s because most of the above measurements are based around the idea of quality. And for 'best' we tend to assume high quality equals better. But what if the success metrics were, instead, speed and cost?

In both speed and cost, the AI tools are already unbeatable. Imagine a business wanted multiple articles and content for some somewhat throwaway search engine optimization or PR activity. It’s simple: the AI wins. I spent hours writing my article; the AI needed less than 10 minutes, and most of that was the team organizing what was needed.

On a cost/benefit analysis there’s only one winner. So we're at 2-1. The robots are coming back.

The software can also be used as an incredible thought-starter engine. At Nucco, we’ve already seen huge value in simply using prompts to begin conceptual thinking.

So that’s maybe 2-2. And it’s not even half-time yet.

But more isn’t necessarily better

I’ll leave you with a final thought; one I’ve heard multiple other commentators consider.

What if we all use AI to create reams and reams of words, and video upon video of content, and none of it is great or good?

The internet is already drowning in subpar clickbait. A 'more of the same' approach is not necessarily solving the problems our industry faces (like relevance, cut-through and engagement).

Maybe we humans can still rally in the second half, because if there’s one thing we still have, it’s imagination and the ability to create the unexpected. C’mon AI, what’s your prompt for that?

Look out for our human v AI test articles. Let’s see who’s best.