Not surprisingly, we’ve had a lot of conversations lately with clients about artificial intelligence and large language models like Bard and ChatGPT. These conversations have also begun to pop up in chats with prospective clients. When part of your business is creating content, and the world is on fire with excitement around ways to auto-generate large quantities of content, that’s not a shock.
What is starting to become apparent though, is that outside of anyone under twenty who’s already used ChatGPT in an effort to write a school assignment, many people don’t have a solid understanding of what these tools can (and cannot) do.
In this piece, we’ll explore:
- The myth of “AI” - what are the limitations of tools like ChatGPT?
- A Realistic Look at AI and Content Creation
- The AI Hype and Very Real Issue of Content Saturation
- The Rise of the Individual Perspective
- How Shifting Content will Challenge Marketers
The Myth of “AI” - What Are the Limitations of Tools Like ChatGPT?
First, we’re going to excuse ourselves as we tiptoe right into the pothole that most people are already standing in - the one where we refer to all the generative content tools available as “AI.” For our purposes, it’s an easy moniker, one most people accept to encompass tools like Bard and ChatGPT, and it works well enough for our purposes. That said, the term “AI” does not really describe these tools. There’s plenty written about large language models and the tools built on them, so that can be delved into elsewhere.
What we’ve begun to understand is that most people who haven’t spent much time working with these tools simply accept the idea that AI can write all the content they need written and that this will be done quickly, cheaply, and without much heavy lifting required by any mere humans.
Hopefully, if you believe that, you’re beginning to suspect it’s not true.
A Realistic Look at AI and Content Creation
Like most things in technology, it’s not that simple. AI has its strengths and can regurgitate a large amount of readily available factual information (up to September 2021, as of this writing). That makes research into general topics straightforward, and the LLMs we’ve used have been good at organizing outputs in an understandable way, which explains why our educational system has some concerns about this technology. For straightforward writing assignments, these tools perform well.
Our issue with the outputs we’ve generated lies in tone — how the information is presented. We've learned that even when we’ve spent months “training” each of our ChaptGPT “channels,” focusing on teaching it the ins and outs of product and corporate voice, and feeding it examples of the kinds of perspective we share, it cannot effectively reproduce these things with new content. In fact, over time, the tool actually forgets what we’ve taught it and we end up beginning again, due to the token limits inherent in the way these models operate.
As a result of this alone, it’s unlikely that we’ll ever plug client requirements into an AI tool and walk away. (TL;DR: AI is not replacing us any time soon.)
The AI Hype and Very Real Issue of Content Saturation
AI is all the rage in the startup community and venture investment circles. Every piece of technology is expected to have a path to AI application.
What does that mean for content creation? Let’s look specifically at two kinds of content:
- Social Posts on LinkedIn (this could be true on other social platforms, but we focus on B2B)
Over-utilization of these tools has led rapidly to a noticeable shift in the kinds of content flooding both channels. The mantra for both used to be "post consistently.” And consistency still matters (for now, at least), but since these tools allow users to create an enormous volume of content quickly, consistency is no longer an obstacle.
The obvious issue is that posting consistently is only one piece of the equation if you seek to create content that pushes the needle for your company in terms of revenue. If you’ve been on LinkedIn recently, you’ve most likely noticed a shift in the type of content evolving there.
- Salesy and almost aggressive
- Bland and repetitive
- Mediocre and replete with common sense disguised as useful tips
We are beginning to talk to clients about posting less frequently, but focusing on more impactful content that cuts through the clutter.
The Rise of the Individual Perspective
We are really focused on the rise of opinionated content, because we think it will reshape the landscape of professional platforms like LinkedIn and blogs. Instead of content that is directive and full of bold claims without much substantiation, we anticipate that after hitting a saturation point, platforms will begin to prioritize more thoughtful, reflective pieces that bear the stamp of an executive's thoughts or a company's vision. It’s possible this shift has already begun as LinkedIn influencers dart about worrying over the latest algorithm shifts.
The posts we’re watching now aren't pieces that can be quickly skimmed through, but require readers to sit down, engage with the material, and digest it – much like a nutritious, home-cooked meal as compared to a fast-food snack. That's not to say that snackable content doesn't have a place anymore, but it’s simply that, with the amount of it available, we're seeing a shift towards wanting content with more depth and substance.
The blogs that are going to prevail will be longer-form thought pieces. They're not just step-by-step guides or lists, but rich narratives filled with unique insights, perspectives, and stories. They might be an executive's opinion, a company's outlook on a current trend, or a thought-provoking piece about why people need to change their habits. This is where the real uniqueness of the brand comes to life and can't be easily replicated by AI.
Shifting Content will Challenge Marketers
This shift towards focusing on quality and a possible drop in quantity of content might require some training and recalibration of expectations for executives who focus on deliverables alone. But the potential rewards – content that genuinely resonates with audiences and sets a company apart from its competition – is worth the effort.
To be clear, we will certainly be using AI tools for the efficiencies they can provide. Will we be using them to write content? No. For all the reasons we highlighted above. A lack of lived experience and emotion mean that as they exist today, these tools will never parallel a good writer’s ability to evoke feelings in readers.
That said, we use these tools for a variety of tasks around the writing process, but right now, we’re not worried about losing our jobs.