Unless you have been living under a rock, with ear muffs on, blindfolded, on another planet, you will probably have heard of a little thing called ChatGPT — powered by GPT-3, and how content and copywriters may (or may not!) need to start looking for a new career.
I decided to put it through its paces for subject matter that is most relevant to my business - B2B fintech.
But first, a little about GPT-3.
What is GPT-3?
From TechTarget:
GPT-3, or the third generation Generative Pre-trained Transformer, is a neural network machine learning model trained using internet data to generate any type of text. Developed by OpenAI, it requires a small amount of input text to generate large volumes of relevant and sophisticated machine-generated text.
GPT-3's deep learning neural network is a model with over 175 billion machine learning parameters. To put things into scale, the largest trained language model before GPT-3 was Microsoft's Turing NLG model, which had 10 billion parameters. As of early 2021, GPT-3 is the largest neural network ever produced. As a result, GPT-3 is better than any prior model for producing text that is convincing enough to seem like a human could have written it.
At The Content Flywheel, we have tried and even reviewed other AI writing tools, and not been overly impressed. But none have had the hype of GPT-3, so I was curious if the reality matched up.
Some examples of GPT-3 for fintech copywriting
For the purposes of my experiment, I decided to start with open banking. I was curious to see how much of the context around open banking — a new-ish, niche-ish field with its own terminology — that GPT-3 would be able to understand.
For the interest of my fellow writers and marketers, I have also assigned a score for each query on a scale of 0-10. This is a subjective score based on how practical and useful I think the output would be for a copywriter writing on the topic of the query I put in, but it is NOT a reflection of how amazing or interesting I think the technology is outside of this use case or in an abstract sense.
Test 1: Open banking payments v credit card payments
First request:
explain why open banking payments are better than credit card paymentsResult:
Open banking payments are better than credit card payments because they offer greater security and transparency. With open banking payments, users have access to real-time data on their finances, providing them with greater visibility and control. Open banking payments also allow for faster and more secure payments, as the data is encrypted and transmitted securely. Lastly, open banking payments often offer lower fees than credit card payments, saving users money.
For a copywriter, this is mildly terrifying, at first glance.
But once one gets over the shock of a paragraph of structured, apparently well-written argument coming from thin air, people who are familiar with the space will notice a few issues.
First, it isn't clear which "users" have access to real-time financial data. Individuals have access to their bank information at all times, and while it is a principle of open banking that third parties can access real-time financial data, it is not clear if the OpenAI text is referring to these third parties. In addition, while the query is focused on payments, the AI lumps open banking account data and payments as one topic, although they both have their own USPs and use cases. And finally, you could argue that open banking payments are secure since data is transmitted securely, but the real reason it is secure is because you authenticate within your banking app, and there are no third parties in the payment flow.
In other words, it looks like it was written by someone who had clearly done some reading, but was not across all the details. Usability score: 4/10.
Test 2: open banking payments v cc payments in detail for a blog post
Second test: a blog post on the same topic. Query:
explain in detail why open banking payments are better than credit card payments for a blog postResponse:
Open banking payments have become a popular alternative to traditional credit card payments. They offer many advantages that make them more attractive than credit card payments.
First, open banking payments are more secure than credit card payments. With open banking, you don’t have to provide your credit card details to the merchant. Instead, you authorize the transaction using your bank’s secure authentication system. This means that your personal details are not shared with the merchant, reducing the risk of fraud and data theft.
Second, open banking payments are faster than credit card payments. With open banking, payments are processed instantly and securely. This reduces the time it takes to transfer funds, making it much easier to pay for items quickly and conveniently.
Third, open banking payments are more cost-effective than credit card payments. Credit card payments often incur fees, such as transaction fees and foreign exchange fees, which can add up quickly. With open banking, these fees are eliminated, allowing for more cost-effective payments.
Finally, open banking payments are more convenient than credit card payments. Many banks offer open banking integration, meaning that you can use your bank’s secure login to pay for items instead of having to enter your credit card details manually.
Ok, this is interesting. On this post, the AI picked up that shoppers authorize the transaction using your bank’s secure authentication system — a key point that was not mentioned in the first. Personal details not being shared with the merchant apparently reducing the risk of fraud doesn't make sense, however.
The second and third points are largely accurate, although lacking some detail and also some important caveats. But the fourth point again seems to confuse the point of view — seemingly talking to merchants at first but then repeating the same point about secure authentication made in the first point. A final quibble from a usability perspective — while my query mentions a blog post, this is more like an outline than something substantive. Score: 4/10
Test 3: Open banking definition for an FAQ page
So my first two tests have been pretty demanding. I was curious to see if GPT-3 could help with a slightly more mundane — but still important — task for a fintech copywriter: FAQ pages. Query:
explain what open banking is for my website FAQResponse:
Open banking is a system that allows customers to securely share their financial data with third-party providers. This allows customers to make better-informed decisions about their finances and to access new and innovative financial services. Open banking also gives customers more control over their financial data, allowing them to decide how and when it is shared. This can help to improve the security of customers' financial data and reduce the risk of fraud.Ok, this is not bad, as far as it goes. It does completely miss the payments side of open banking, which is quite a glaring omission. But it isn't a bad starting place for a knowledgeable copywriter to start building on a more thorough and well-rounded definition. Usability score: 6/10.
The verdict: GPT-3 is a new tool for the writer's toolbox, but not a replacement for the writer — it (sort of) says so itself!
These test queries may look highly specific and in-depth. But in actual fact, there is far more specific content that your copywriters work on every single day that no AI is close to being able to produce copy for. These include content types such as case studies, which are specific to you and your customer, product launches, which are unique to your business, and so on. For that reason, copywriters, you should be confident that you are still needed. And marketers, hold your copywriters close, for theirs is a skill that AI can not yet compete with.
However, when asked:
Will ChatGPT make B2B tech copywriters redundant?The answer (my emphasis added):
No, ChatGPT is not designed to replace B2B tech copywriters completely. ChatGPT is designed to be used as an additional tool or resource to help copywriters create content that is more relevant and personalized for their clients. It is not meant to replace the expertise and creativity of a professional copywriter.The words "completely" and "not meant" are doing a lot of work in this answer, but without being very reassuring that the ambitions behind it are particularly modest. 😲
Nonetheless, it is already both exhilarating and a little unnerving to see how fast AI copywriting has progressed. For the moment, GPT-3 can probably be a handy writing tool to speed up, fine-tune, and generally improve your overall content efforts for many generic pieces of content, alongside trusted tools such as Grammarly, Hemingway, AnswerThePublic, and so on.