AI Content Writing: Our Take on ChatGPT-3
5 SECOND SUMMARY
- AI is here and it’s only going to get smarter. We’re on the cusp of a new industrial revolution where humans work side by side with AI tools to create better work.
- Positives: AI technology is another useful tool in our toolbox, helping marketers get things done faster, and taking away some of the ‘grunt work’ of our jobs.
- Negatives: so far, it still takes time to ‘drive’ tools like ChatGPT-3, with a significant amount of work to turn generic text into something with personality, expertise and (importantly) accuracy. ChatGPT-3 can only pull older information that already exists too, meaning no new perspectives and no capability to write something like a case study.
- As with any new technology, there will be haters and evangelists, but we’ve tried to present a balanced view here.
- The most important thing to remember: so much marketing energy is wasted on figuring out how to hack the system instead of just doing good work and being laser-focused on your customers.
There’s a huge amount of chatter in the marketing sector right now about AI content writing tools.
Here’s our take on the good, the bad, the weird, and the future of content marketing.
If you’ve been on LinkedIn the past few weeks, you’ve likely seen marketers and agencies waxing lyrical or freaking out about ChatGPT-3, an AI chatbot that some are hailing as the next best thing in the marketer’s toolkit.
What is ChatGPT-3?
ChatGPT is a publicly available prototype of an artificial intelligence deep learning model trained on the data currently available on the internet. It’s used for answering questions, summarising text, machine translation and code generation.
Specialising in dialogue, it combines some of the functionality of Google Search with tools like Alexa or Siri – basically, it scrapes the internet for the most likely answer to your question and returns the result in a conversational way.
And, because you can pose a question and sometimes get a few paragraphs that offer a ‘good enough’ result, it appears on the face of things that your content writer is about to become defunct.
If you can just plug in a question and get a result in 10 seconds, is it time to ditch your content writer and save some coin?
Well… not necessarily.
Will we see a big upswing in fairly bland, run-of-the-mill content being pumped out by some businesses? Probably. Will entry-level writers or those that have limited creative and critical thinking skills be anxious about their jobs? Most likely. Will AI simply take the place of all writers? We don’t think so.
To offer a balanced view of what people are saying about ChatGPT-3, we took to the world’s soapbox (LinkedIn) and captured a range of differing opinions.
Here’s what we found.
First: the bad
Some marketers bemoaned the fact that tools like this were about to suck what little personality was left out of the content marketing landscape, producing surface-level content made by robots, for robots.
Drew from KlientBoost, Travis from PandaDoc and Nick from IMPACT are not fans:
Writer Luke O’Neill has put the spotlight on a comical situation that could unfold, where we see robots (AI) writing content for other robots (Google’s crawlers). This is likely to culminate in marketers having to measure the success of promoting products to an audience that doesn’t actually exist.
This would truly be a bizzare situation.
Matt Aberline tried a few queries and found some ‘good enough’ answers, challenging his network to think about how technologies like this will impact everyone’s careers.
GPT-4 is already on the horizon and is likely to be capable of a lot more choices of “next word” or “next sentence” or “next emotion” based on the context. This means it will be more similar to actual human thinking than its predecessor.
The indifferent:
Several people have reminded us of the fact that this technology is actually not new.
Although it seems like AI has burst onto the scene with full force over the past few weeks, in reality, it is simply another slightly refined iteration of tools like Google voice search which has been around for several years now.
In fact, Google has been doubling down on AI for the past six years, recently announcing its own experimental conversational AI service called Bard.
Currently in beta, early information from Google indicates that we may be seeing fewer featured snippets as the search giant aims to succinctly answer people’s questions (as seen in this image on the right).
From Google:
“AI can be helpful in these moments, synthesizing insights for questions where there’s no one right answer. Soon, you’ll see AI-powered features in Search that distil complex information and multiple perspectives into easy-to-digest formats, so you can quickly understand the big picture and learn more from the web: whether that’s seeking out additional perspectives, like blogs from people who play both piano and guitar, or going deeper on a related topic, like steps to get started as a beginner. These new AI features will begin rolling out on Google Search soon.” Source
It will be interesting to see how this plays out and whether the Holy Grail of ‘position 1’ becomes less critical (or even ceases to exist) as more and more complex questions are answered directly from the search results page.
As for GPT-3, it was first announced in May 2020 and some have hailed it (and its future iterations) to be the ‘death of Google’, replacing the way the world searches for information. However, conversational technology has been around for a while now and is part of our lives already. Tools like website chatbots are widely used, answering simple questions and filtering more complex queries through to human operators.
So will Generative Pre-trained Transformer (GPT) models replace Google or just add a large dose of healthy competition?
Mark Williams-Cook from UK digital agency Candour doesn’t believe Google is going away any time soon.
Mark has put this very succinctly in this visual, and I have to agree with him.
I believe we’ll see a huge uptake of people using these tools; then realise that crappy content is still crappy, we’re just pumping it out faster; then take a step back and start to thoughtfully and sustainably integrated AI tools into our existing workflows.
The partnership view:
Several people have described how AI functionality can become another tool in our toolbox, helping marketers get things done faster and taking away some of the ‘grunt work’ of our jobs.
This will require a shift in the way we work, from being ‘doers’ to being in the driver’s seat, guiding AI tools to our advantage while our feet are firmly planted in real-world business strategy.
As Jonathon Boymal notes below, this approach will need marketers to bring their creativity and strategy skills to the fore, asking the right questions and providing the right inputs, then using any AI-generated content as a rough draft only.
Then, using their creativity and understanding of the nuances of the business, marketers can mould and refine the rough output to produce high-quality finished work.
Brian Miller cautions that understanding the context we are working in is something that AI has yet to master. The human touch is essential to add context, relevance, and tie the information back to what we inherently know about our customers.
Whilst ‘users’ might be predictable, logical, quantifiable and happy with binary choices, this doesn’t hold true for real human behaviour.
People are messy, complex, unpredictable, constantly evolving and shifting in their opinions, and often resistant to solutions.
We take action on gut feel, we are indebted to social exchanges, we are stubborn and contrarian. Uncovering what makes people tick is the driving force behind almost ever marketer I know, and being able to bring this deep knowledge to the table is key to producing content that hits the mark.
If marketing only required logic, our jobs would be SO much easier!
Copywriter Glenn Peters has beautifully summed up the value that good writers provide – and ‘typing out words’ is an incredibly small part of what we do.
Love it:
And of course, to balance all this, we have those that are simply loving this technology and are fully on board with the potential it will bring to all kinds of businesses and industries.
Brendon Ross is all-in, and Ted Strazimiri has successfully used AI to cut down his coding time and boost productivity (yep – ChatGPT-3 writes code as well).
And for the mathematically challenged like myself, tools like AI bots that deliver Excel formulas from a plain-English sentence are destined to be incredibly helpful (“Hey Excel – give me the sum of column A when column B contains the word “marketing” and column C is today’s date”).
We’d be remiss if we didn’t mention how much fun people are having with AI too.
Bronagh Brbich asked for an explanation of the NSW Planning system in the style of Dr Seuss:
Jonathon Mann from Fundamentals Australia asked ChatGPT-3 to write a children’s story about IEC 61850, the international standard for communication protocols for electronic devices at electrical substations.
Key takeaways:
Ultimately, AI is here and it’s only going to get smarter. This could signal the dawn of a new industrial revolution where humans work side by side with AI tools to create better work. Teaming up with robotic assistants will change the way we create, eventually moving writers into the role of an orchestra conductor, where we guide, enhance, refine and add colour and context to raw outputs to help content stand out and ensure it has personality.
It’s also important to note that currently, tools like ChatGPT-3 can only pull older information that already exists: so if everyone used AI tools in excess, everything would sound the same, and there would be no new opinions, angles and thoughts. Generic content is a big enough problem already – do we really need more?
Another key point to note is that using AI tools still takes time. If you’re outsourcing your content marketing because you are strapped for time, you don’t want to sack your writer and then spend loads of time driving an AI tool, editing, finessing, rewriting, and removing incorrect dates and facts (which seems to be happening often). From my own experiments, there is a significant amount of rewriting to do, and it hasn’t actually saved me any time as yet.
My personal take as a content specialist is to be aware and ready for shifts. Hopefully, AI will help eliminate mediocrity and force us all to lift our game. It’s likely to increase the speed of creating first drafts, but on the other hand, it will slow down the editing process.
What it can’t offer is expertise, unique perspectives, and the kind of battle-hardened knowledge that comes from spending decades in your niche industry.
Gary Wollenhaupt sums it up perfectly below:
“Lots of marketing/search energy is wasted on figuring out how to hack the system vs. being really freaking good at what you do for customers.”
Keep your eyes on the prize and always come back to ‘what do my customers want?’ – this is always where marketers should be aiming, now and in the future.
Get help with your content marketing
