ChatGPT. It’s all we hear about these days. As an artificial intelligence chatbot developed by OpenAI, it has the potential to be a major disruptor, reshaping content creation as we know it. As of December, ChatGPT has attracted more than 100 million monthly active users. So many, in fact, it has to periodically ask new sign-ups to join a waitlist, and most recently, only paid subscribers have full access to the platform.
As prolific digital marketing content creators for our clients, we are just as curious as everyone else. From developer code and eCommerce personalized assistants for shopping recommendations to short-form marketing content and social posts, ChatGPT and other AI-based solutions can generate just about any type of content a user asks of it.
Companies and people are jumping on the ChatGPT bandwagon and finding ways to integrate it into the user experience and life itself. It’s opening up thought-provoking conversations about AI’s role, its benefits and risks and whether it brings us humans that much closer to being replaced by the technology we created.
ChatGPT: Friend or Foe?
There are many use cases for AI content, but user beware. It isn’t as cut and dry as you may think. As with most things, there are pros and cons. And with AI, many of the cons have yet to reveal themselves.
ChatGPT is, for now, the frontrunner in AI-assisted writing. At its best, it can save companies money, speeds up writing, and creates surprisingly coherent content on just about anything you can think of in just seconds. Similar to how we responded to the genesis of the internet, the results from a search for information on ChatGPT are at first shocking. The first time you use it, you’ll either shake your head in amazement or verbally express your disbelief—“How did it do that?”
Some are already realizing AI assistants can actually help authors write better or at least make them more efficient. For example, generating blog ideas is time-consuming, particularly if you’re writing multiple blogs every month on similar topics. ChatGPT can help generate fresh ideas and even provide a topic outline given the right prompts.
One of the key components of SEO writing is to include reputable, relevant links to third-party sources to back up claims, provide discussion points and lend credibility. A user can ask ChatGPT to find a source on a topic, saving the writer research time. No guarantee the sources returned will be non-competitors or reputable, however.
The AI assistant can also help with shorter-form content, like social media posts. If your brand consistently posts, using ChatGPT can save you time and constantly create new content.
While there is great potential for AI and viable use cases for such technology, it has a darker side. Perhaps we are questioning it so early because we didn’t necessarily see the canary in the coal mine when it came to the internet or social media. Lesson learned. Great for some things, but can be used for nefarious purposes or put the user at risk as well. And that’s what so many are afraid of with AI-based assistants. There have yet to be any standards or governance, rules or protocols to guide users. It’s sort of the wild west right now.
Just days ago, The Wall Street Journal reported that Apple has delayed approval of an email-app update that uses AI-generated language tools to automate email content creation due to concerns that the request results could present inappropriate material to children using the app. With so much focus on social media right now, it’s no wonder Apple is nervous about the impact this technology could have on children and teens.
School teachers and college professors are grappling with the idea of ChatGPT writing as well. Questions swirl around whether AI will inhibit or enhance learning. Is it unfair? Can students learn a subject and how to write well if they’re relying on ChatGPT to write it for them? And what about plagiarism? Is it akin to Grammarly and could help students become better writers? EdWeek says that until there is consensus on how to handle this innovation, many educators are planning on requiring students to write essays by hand, old-school style sans technology.
These are just a couple of examples of how AI is stirring the pot, and these types of conversations are a good thing. We need to think through AI and learn how to control it before it controls us. But what most people fear the most about ChatGPT is that it will replace humans in the workplace. What was once considered too far off to worry about is here. And it’s only gaining steam. A Resume Builder survey found that 49% of companies use ChatGPT, and 48% say ChatGPT has replaced workers.
As a digital marketing agency that generates hundreds of pieces of human-generated content per month for our clients, do we use ChatGPT, or do we plan on letting our content team go and let ChatGPT generate our content? Absolutely not—on both counts.
ChatGPT Can Be a Human Enabler, Not a Replacement
CBS News notes that while the majority of jobs may be secure now, AI will only get smarter and better, and people must adapt. Interviews with multiple experts and ChatGPT users reveal a common sentiment—that AI can remove mundane tasks and create new jobs and greater job satisfaction and greater productivity. It even goes so far as to predict AI will drive up wages. Much like phone banks have been replaced, workers will need to evolve, learning new skills and using AI to do their job better, faster and more efficiently.
After experimenting with ChatGPT, it was clear that the quality, accuracy and effectiveness of its writing results could not compare with skilled human writers. Here are a few observations comparing humans with ChatGPT:
Writing quality was subpar compared to skilled humans
ChatGPT can respond with content, but that content is typically bland, textbook-style, short-form writing without much personality. There’s a reason for this—it gets its data from Wikipedia, webtexts and articles on the internet. It isn’t creating new content as much as it is collating existing content together, and each source has its own style and tone. By diluting those multiple styles into a singular style, ChatCPT can present content that is informative but undistinctive.
If you’re just presenting information, a definition or a direct explanation, ChatGPT may work fine. But when it comes to blogs, for instance, companies need to use them to differentiate themselves, revealing their brand identity to build rapport with the audience. Think of ChatGPT content as a doctor’s bedside manner. You may get the information you need, but it’s delivered in a way that can be uninspiring and dull.
As mentioned above, some question whether AI-generated content is like Grammarly, the online editor that finds grammatical and punctuation errors. Grammarly is awesome, and users can actually learn grammar, punctuation, and sentence structure and how to write clearer, easier-to-read content using it. Can the same be said for ChatGPT? Perhaps.
But as a proficient user of Grammarly, I know well that it is not always correct. In fact, I catch quite a few Grammarly errors that someone who isn’t an experienced writer may miss. Even with Grammarly’s help, I have to double-check every flagged “error” and re-read the content to make sure Grammarly didn’t miss anything. In both cases, it is rare that I don’t discover a Grammarly mistake. The same goes for ChatGPT. It is not capable of delivering perfect content that needs no human editing.
Content accuracy was questionable
AI is not perfect. Type in a blog topic to ChatGPT, and you may get completely irrelevant, even weird content that is downright wrong. ZDNET says, “At this point, it’s worth remembering how tools like ChatGPT work: they put words together in a form that they think is statistically valid, but they don’t know if what they are saying is true or accurate. That means you might find invented facts or details or other oddities.”
The Guardian says it this way: “The bot doesn’t work perfectly. It has a tendency to ‘hallucinate’ facts that aren’t strictly true, which technology analyst Benedict Evans described as ‘like an undergraduate confidently answering a question for which it didn’t attend any lectures. It looks like a confident bullshitter that can write very convincing nonsense.’”
With this in mind, all proposed facts by ChatGPT must be researched and validated. Even definitions, claims and statements that appear to be accurate are deceptively wrong, which then reduces or eliminates any productivity and efficiency gains. And because a human didn’t create the content, finding data to substantiate claims can take quite a bit of research time.
SEO techniques were nearly non-existent
Google periodically releases algorithms to its page rankings, giving marketers insight into the tactics they should use to ensure their content is as high up on page one as possible to get the most views and clicks. Even the best-written human blog can fail to appear on the first page if the right SEO tactics weren’t used on the front and back ends. ChatGPT isn’t considering SEO.
For instance, in the newest Google algorithm release, one of the top requirements is for the content to be original, written by an “enthusiast who demonstrably knows the topic well” and cites recent, relevant sources. Google specifically recommends that the content should be factual (which we just covered) and should be so trustworthy that readers would feel comfortable trusting the content for issues relating to their money or their life. Pretty high standards.
ChatGPT cannot check all of these boxes, nor can it produce long-form content without significant user input. For blogs to rank, experts recommend the ideal blog post length to be 2,100- 2,400 words, with 2,500 words being the preferred length to generate leads. Why? Longer content goes more into depth and has more real estate to cover a broader range of topics, which is precisely what Google is looking for when it is assuming what a bunch of searchers may want to know. It also reinforces your brand’s topical and industry authority.
For ChatGPT to create this much content for one article, the user still needs to continually prompt it with new topics and ensure content flow, accuracy and consistency without repetition. Unfortunately, at these lengthier word counts, we found the articles began to repeat themselves instead of generating new ideas in a coherent way. Editing for this and tweaking it to include on-brand personality and tone took nearly as long as generating the content ourselves, and it still didn’t seem as high quality as what we produce every day.
The risk of plagiarism is too high
If there is one cardinal rule of writing, it’s not to plagiarize. Ever. Most everything else is up to creative freedom. Unless the ChatGPT user includes a disclaimer that the content or portions of the content were generated by ChatGPT, it is considered plagiarism. Sadly, many people do not include any reference to the AI assistant and pass the work off as their own; thus, the problems schools are facing.
Schools approach plagiarism a bit differently than companies, where plagiarism is not only unethical but impacts SEO, can result in a massive hit to brand reputation, and, if detected, can lead to legal consequences.
Even OpenAI, the company that created ChatGPT, recognizes the problem of plagiarism, releasing a new tool called classifier that is supposed to identify content that has been authored by AI and passed off as human. But it, too, is not yet perfected, only identifying 26% of AI-written text.
Beyond not crediting AI, there is a debate as to whether ChatGPT returns plagiarized results. Because it scans the internet to find its content, there is a risk that the AI captures text and reposts it in its own responses. It’s important to note that plagiarism can appear in more ways than obviously copying a sentence, paragraph or entire article verbatim.
To quote the National Association of Sales Professionals, there are various forms of plagiarism, all of which can have negative consequences if the source(s) are not properly credited:
- Patchwork plagiarism reflects the act of plagiarism in which an entity picks writings from different sources and mix and match them to create new writing and publish it under its name.
- Verbatim plagiarism is the type of plagiarism where plagiarists pick up selected words or sentences from someone else’s writing and include it in their content.
- Paraphrased plagiarism means the type of plagiarism that involves rephrased words reflecting an idea or topic that was once discussed by someone else but in different words.
- Direct plagiarism is the act of directly picking the entire writing articulated by someone else and publishing it under your name.
- Self-plagiarism is yet another type of plagiarism in which an entity includes its prior-written work in the new writing or presents old writing as the new one.
- Accidental plagiarism is the type of plagiarism in which you may find duplicated or influenced writing to some other sources. It mostly occurs because of common words.
Yes, AI-generated content can save a marketer time, but it still requires human intervention to ensure it is not plagiarized, is accurate, of high quality, and its tone and style are on-brand.
Can Companies Rely Soley on ChatGPT to Generate Marketing Content?
ChatGPT is an exciting new toy that has plenty of use cases, but it still has a way to go before it can completely replace a human. Our humans at Zilker Partners remain committed to writing 100% unique content that hits all of the marks for being a well-written, accurate piece of work that considers our client’s individual value propositions, tone and style.
While we may one day use ChatGPT as a resource to help us generate unique blog ideas, we still rely on our experienced writers to research and deliver content that persuades, informs, educates and delights readers as it moves them further down the road in their buyer’s journey.
Companies hoping to save marketing dollars by using ChatGPT can likely find ways to incorporate it into their processes; however, the technology is not capable of delivering the nuanced content in certain assets. As many experts are discussing, ChatGPT and other AI content platforms are not going anywhere and are only getting smarter, but humans play a vital role in ensuring that content meets certain standards.
Skilled writers are necessary to tell the bot what to do in a way that will return the best results and edit the content with a critical eye to ensure the highest quality content. AI technology may help us be more efficient and productive, but in the end, it still needs us.
We’ve heard it said that AI platforms are akin to robotic surgery. The robots aren’t replacing surgeons, only making them more precise and efficient. Writers can learn to use AI technology to work in their favor without relying on it to do all of the work for them. There will be a learning curve for marketing content writers to optimize its usage, but there is no doubt that ChatGPT and similar platforms are here to stay. Contact us for the highest-quality, human-written digital marketing content.