By now, most everyone is aware of AI (Artificial Intelligence) and the remarkable progress that’s been demonstrated lately by generative AI tools like ChatGPT and Google’s offering, Bard. You may have even used one of these tools to assist you in the course of your own work. But is it a good idea?
Since we’re still in the relative early stages of widespread adoption of AI, the answer isn’t exactly cut and dried. Let’s take a look at some of the potential pitfalls in using AI to perform marketing functions like generating taglines, product descriptions, blog posts and other things you might produce in the course of your marketing work.
Before we really dive in, it might be useful to take a look at exactly what generative AI is and how it works. And who (what?) is more suited to lay that out than AI itself?
Here’s what ChatGPT has to say about the subject:
“Generative AI is a type of artificial intelligence that can create new content, such as images, text, and music, without human intervention. It works by using deep learning algorithms to analyze patterns and relationships in large datasets of existing content. Once the model has learned these patterns, it can generate new content by using this knowledge to predict and create new data points. Generative AI has the potential to revolutionize many industries by automating content creation and generating new insights from existing data.”
So, we’ve already used AI to write some of this blog post. Is that kosher?
Since it’s clearly indicated that AI – rather than a human – wrote that paragraph, it would be difficult to argue that it’s not. But what about other cases where someone is using AI behind the scenes to generate content and not disclosing that? This is where things get dicey, to say the least.
Richard Rimer, an Atlanta-based Intellectual Property attorney and member of the id8 Collective, explains it this way:
“Your knowledge is the sum of everything that you’ve ever been exposed to in your life: the books, the music, the art, and whatever else you’ve been exposed to. You have at least a vague notion of where you’ve got all that information and when you’re being creative, you’ll draw from that necessarily. You probably didn’t create most English words you use – you’re borrowing from everyone.”
The hallmark of creativity is to synthesize that information and present thoughts, ideas, or even art in a way that they’ve never been put together – or at least used – before. Rimer says that five or ten years ago, all you needed to do was make your best effort to be unique and original and not steal from anyone intentionally – or infringe on their copyright.
One of the biggest potential issues with AI is that the words that it presents in response to your prompt all necessarily came from somewhere else. Remember, even AI itself admits it draws from “large datasets of existing content.” AI generally doesn’t present any annotation to the content it produces, so you have no idea what the source of the words is or how the AI tool put it all together.
“[The AI] could have gotten all of the information from a single blog,” Rimer says. “It’s unlikely, but possible. Or it might have gotten key sentences that you choose to use from a blog post. If you quote those key sentences, you would become an infringer of the blog post unknowingly.”
There’s another potential problem with all of this on the flip side.
Copyright laws vary around the world, Rimer says, generally the rule is that unless an actual human writes something, it can’t be copyrighted. If you borrow substantially from a tool like ChatGPT in your marketing materials, there’s nothing to stop one of your competitors from copying you in turn since you have no ownership of the words you’ve put out there.
Legal issues aside, using AI to perform marketing functions that would traditionally be assigned to a human presents some more immediate challenges. We touched on creativity, and that is something that AI is utterly incapable of at this point. Remember, all it can do is piece together already existing content in a way it thinks (or, more correctly, processes according to its algorithms) makes sense.
AI can generate text that is grammatically sound and technically correct, but it can’t necessarily create copy that is truly engaging, resonates emotionally, and is true to your brand’s voice. Remember, it’s a computer – it doesn’t have emotions like Big Hero 6 or even really understand what emotions are. All it can do is describe them based on what people – who actually experience them – have said or written.
Truly great marketing copy involves things like storytelling, unique perspectives, and emotional (there’s that word again!) appeals – all of which are difficult, if not impossible, for AI to replicate.
And good luck trying to trick your customers: We all know that copy that is formulaic, generic, and impersonal is a recipe for failure – and potentially even damaging to your brand.
All of this is not to say that you should never use AI in your marketing work. As it stands, it certainly has a place in idea generation and other jumping-off points – but that is a subject that deserves to be tackled on its own, in another post.