310
310
Read it in 11 minutes
Table of contents
Drawing more than a million users after it was first launched in November 2022, talk of Open AI’s ChatGPT tool has been taking over my LinkedIn feed. From jokes to poetry, Excel formulas, code checking, or even practising for a job interview, there seems to be no end to its talents.
You can ask it to do all sorts of tasks, just for fun…
And if you’re not happy with it you can ask again:
Of course, content writers are particularly buzzing to explore how ChatGPT’s capabilities can help produce stand out content for their target audience. Some content marketers were even starting to fear that AI might be coming for their jobs – although it’s clear that we’re still a long way off that. To understand its potential, let’s start with how it works…
To answer this question, I thought I’d be very meta about this and ask ChatGPT to explain itself… Here’s what it said:
ChatGPT is, in the words of the New York Times, “the best artificial intelligence chatbot ever released to the general public.”
It uses a type of language model based on deep learning techniques to serve up human-like text in a chat format (with a personal tone). It’s trained on large volumes of data and it learns statistical patterns and relationships between words and phrases – so that it can understand words in context. Once trained, the model can generate new text that’s similar in style and context to the training data.
Amazon, Apple and Google use large language models trained to support products like Alexa, Siri and Nest. However, these have been kept largely under wraps and so far have been used to provide information and execute tasks based on commands. GhatGPT is more conversational, designed to engage in back-and forth conversation. Unlike Siri or Alexa though, it relies solely on a text-based interface.
As OpenAI highlights on their home page: “We’ve trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.”
ChatGPT’s encyclopaedic knowledge comes from a wide range of text data – such as books, articles and websites. However, it’s important to remember that it doesn’t speak with sentience (even if it sometimes appears to!) and it doesn’t have any understanding or consciousness of the world outside of the data it’s been trained on.
Here’s what ChatGPT has to say about its use as a content creator:
What struck me here was its awareness that whatever it produces needs to be supplemented by human knowledge and expertise; at the very least, it will need some fine tuning.
Don’t make the mistake of thinking ChatGPT content is ready for distribution as is – the editing process needed is often fairly extensive. However, it can still help make the content creation process faster and easier, if used in the right way.
You can use it to:
There are some specific and handy uses that relate to SEO. It can:
The ingredients for creating engaging content remain the same as they have done for years – and ChatGPT doesn’t have the context to be able to bring all of these things together. This is still a job for content marketers.
We know that the best content is built on the foundation of knowing your audience – understanding their goals, challenges and priorities and creating content that speaks to these themes. The best way to get to the heart of your customer’s challenges and wish-list is to interview them and conduct detailed research – something ChatGPT can’t do for you.
We also know that telling a story that speaks to human emotion is important to connect with our readers. However, ChatGPT lacks the context of human emotion – ask it to tell a story and it can feel rather flat. As Robert Rose from the Content Marketing Institute puts it, “You can see that ChatGPT is good (quite good, actually) at constructing sentences that logically flow from one to the next. But there’s no uniquely emotional point of view – or even anything that resembles building a story. Simply put, ChatGPT can do plot – it can tell you what happened. But it’s not great at telling you in a way that makes you feel anything.”
Ultimately, ChatGPT lacks the creativity that humans possess. It can write you a reasonable blog, but it won’t be unique or polished – it also doesn’t have the ability to create all the other elements content marketers use to bring their content to life like images, graphics and gifs.
AI can’t connect to the reader in the way a content marketer can.
Increased use of AI-generated content will mean social media and search will become even more crowded with similar-sounding content. Organisations and individuals that rely too heavily on text generated by ChatGPT will end up producing content that’s repetitive, lacks originality and is less engaging.
As Danny Asling points out, ChatGPT is missing curiosity – it lacks the ability to ask the tough market questions or create something that hasn’t come before. Content marketers that rely solely on what ChatGPT generates will find that their content is the same as everybody else’s, and doesn’t differentiate their brand.
And the ease and swiftness with which ChatGPT produces content (it only takes seconds) means there will just be a lot more content out there to wade through. This will make it harder to cut through the noise and get your customers to take notice of your content.
To counteract trend one, we’ll see a growing number of brands and individual content creators putting effort into creating original insights, producing more creative campaigns, and promoting genuine thought leadership that will be valued by their target audience.
It’s not just about getting eyes on content but ensuring that content is truly useful and thought-provoking for readers…
Content marketers will need to lean more heavily on insights from in-house topic experts (now is the time to speak to that consultant on your team who has 30 years experience in the industry), sales teams and customer success teams – as well as customers themselves – to create something unique.
Original research that shares new perspectives, evolving trends, and customer stories should be at the top of your content creation agenda. And the visual elements of content will also become more important to help it stand out from the crowd.
ChatGPT can be used to generate personalised content at scale, such as emails, social selling messages, and even landing page copy.
However, this kind of personalisation will largely be superficial, which won’t benefit marketers as much as they hope. To truly create a personalised experience that goes beyond FirstName, LastName, CompanyName, JobTitle, content marketers need to understand the challenges of the account, and the key decision makers, and take a creative approach that wows their target audience.
‘Personalisation at scale’ is a bit of an oxymoron; trying to scale personalisation is not really true personalisation. The most compelling campaigns take significant time, resource and research investment. It’s likely that brands will focus their content efforts on accounts with the greatest revenue potential, opting for a one-to-one or one-to-few ABM approach instead.
This issue is very much still up for debate, but some experts have already pointed out ChatGPT’s potential to infringe on intellectual property rights because it’s trained on vast amounts of text data, including books and articles. If this training data includes copyrighted material then ChatGPT’s outputs could infringe on the copyrights of these works.
It’s also worth being aware that it could produce offensive content that could open you up to legal challenge – ChatGPT is up front about this risk, stating that it ‘may occasionally produce harmful instructions or biassed content.’
As the use of ChatGPT continues to go more mainstream, we might see some of these legal implications start to play out in 2023.
Like any tools we suddenly find ourselves using, there are also ethical concerns over how it can be misused – nefarious uses could include plagiarism, cheating at school assignments or using it to write malicious code or email phishing campaigns. One potential solution that’s already been discussed by Open AI (the creators of ChatGPT) is “watermarking” content using an “unnoticeable secret signal” to show that it’s been produced by AI. However, this is yet to be released.
It will be interesting to see how the likes of Google and LinkedIn respond to the widespread use of AI-generated content – ultimately they’re not going to want lots of the same content making their platforms less engaging. Indeed, last year, Google said that AI-generated content was against its guidelines. Will Google, LinkedIn and Facebook evolve their algorithms detect AI-generated content and demote it?
Some people might also start using ChatGPT more often to answer their questions than they use Google – however, there are a couple of issues with this:
Firstly, ChatGPT isn’t particularly good at quoting its sources (meaning writers basing their articles on ChatGPT won’t be sourcing things properly or know where the information came from originally, unlike on Google SERPs).
And secondly, ChatGPT’s short, condensed answers are a very different experience to Google search results. We’ll likely end up reading less content around a particular topic – from fewer sources – and may not end up with a very balanced view, depending on what ChatGPT serves us. Surely, this has the potential to perpetuate biases or narrow people’s point-of-view. This isn’t a big leap to make; the algorithms that power social media platforms have already been liked to the manipulation of peoples views and increased political polarisation, as we saw on Netflix’s hit show, The Social Dilemma.
As examples of these concerns emerge, people may become more cautious and ensure they are not relying on ChatGPT as their single source of truth.
Although ChatGPT is a very impressive tool, it lacks the benefit of human creativity, connection and experience.
But like all the revolutionary inventions that have come before it – typewriters, computers, the internet – ChatGPT will change the content creation process. Being able to use it, and use it well, will help content marketers to be more successful at their jobs. It’s a resource, and a valuable one at that. Just make sure you add your own knowledge, tone and flair! AI tools like ChatGPT are still developing, but by starting to use them now, you’ll be at an advantage as they evolve.
ChatGPT can only get your content creation process so far… content marketers need to take it the rest of the way.
Sign up to receive B2B Search Marketing Insights direct to your inbox
Book a call to discuss your business' search marketing needs
Book a call