Chat Generative Pre-trained Transformer, more commonly known as ChatGPT, has been all over the news lately. Developed by OpenAI, it was released as a prototype in November 2022 and quickly gained national media attention for its depth of knowledge on a wide range of subjects. But ChatGPT’s functions go well beyond just answering queries one might ask of a search engine. It can also rapidly write code, articles, cover letters and resumes. ChatGPT can even tackle more artistic endeavors like writing music and creating graphics and digital art. Some have even used ChatGPT to solicit relationship and investment advice.
With the wide range of ChatGPT capabilities and advancements in AI, many in the media have speculated about the demise of some professions. This Business Insider article list most technology, legal, finance and media industry jobs, along with teachers, traders, accountants and graphic designers, as being at risk of being replaced by AI-powered bots like ChatGPT. So is this a reality, and should we all be polishing up our resumes? Or is it something that will always be just over the horizon, like the flying cars we have been promised for decades? Let’s take a closer look at ChatGPT, how it applies to marketing and if it will be the end for all of us.
CHATGPT IS AMAZING (SORT OF)
Part of the shock and awe factor of ChatGPT is that it seemingly came out of nowhere and seems almost unlimited in what it can do. Yes, most of us have heard about artificial intelligence for a while, but it was always nebulous. We suddenly have a real and breathing example of AI right before our eyes. And what it can do is straight out of a science-fiction movie. The media rightly hyperventilated, as did many that used it. ChatGPT is truly remarkable in terms of what a “machine” can do.
Dive deeper into the details, and that amazement starts to wane. Like with a budget paint job for your car, it might look good from a distance, but the closer you look, the more things fall apart. While ChatGPT is amazing for what an AI can do, most of what it produces falls short of human standards. We can’t attest to its expertise in code or music, but we have experimented with ChatGPT writing articles and press releases along with generating images and graphics. Most of what was written by ChatGPT was very clunky and read as if it was authored by someone with only a basic understanding of English. The images and graphics weren’t much better and sometimes even disturbing, with humans having claw-like hands and distorted faces. We aren’t the only ones to experience this, as CNET paused publishing all AI-generated articles and removed some that contained mistakes and were factually inaccurate.
YOU ARE ONLY AS SMART AS YOUR TEACHERS
A lot of these glitches in the matrix come from how ChatGPT learns. According to its Wikipedia page, ChatGPT uses a mixture of supervised and reinforcement learning. Both involve a human “trainer” interacting with the AI to improve its performance. In supervised training, the AI learns from conversations where the trainer plays the role of ChatGPT and a user. During reinforcement learning, the artificial intelligence responses to a theoretical user are ranked by a trainer, helping ChatGPT learn which ones are better. Microsoft collaborated with the training using its Azure supercomputing infrastructure, explaining ChatGPT’s wide-ranging capabilities.
All of that sounds impressive, but it mainly deals with the conversational abilities of ChatGPT. Use it for other tasks like writing a blog post, and it quickly becomes apparent that ChatGPT is mining the internet for information. That is where things start to fall apart as the internet is not the best “teacher” for many subjects. It can be filled with misinformation, inaccuracies, plagiarized material and other refuse. Currently, AI platforms like ChatGPT that crawl the internet struggle to separate actual facts from the volumes of non-facts online. And ChatGPT’s struggles multiply when there is less information on a subject online, making it produce borderline nonsensical content.
AI only being as good as what it is taught was even a problem before the birth of ChatGPT, as the excellent article “The Foundations of AI Are Riddled With Errors” from Wired explains. It is like going to a low-budget karate school when you have no experience in martial arts. You don’t know if what the instructor is teaching is correct or good because you have nothing else to compare it to. And right now, ChatGPT is going to the lowest-budget “school” of all, the internet, to learn with mixed results.
THOSE DEAD ROBOT EYES
The military was interested in artificial intelligence long before the advent of ChatGPT. Some of the world’s most advanced AIs have been developed by secretive DARPA-backed initiatives. The reason for the interest is simple: swap one of these AI “brains” into something like Boston Dynamic’s somewhat scary and advanced Atlas robot, and you have an emotionless super soldier. It doesn’t have fear, get tired, suffer from low morale or experience the horrors of war. It simply executes the tasks you give it without question or feelings.
There is a flip side to removing emotions from battle, however. Every professional military in the world rewards valor for good reasons. It can turn the tide in battle and pushes individual soldiers to achieve what most would think impossible. Without that very human emotion of courage, countless battles would have been lost throughout history. And anyone who has read Stephen E. Ambrose’s “Band of Brothers” or watched the subsequent series knows that, ultimately, humans will fight hard because they care about their fellow soldiers. An AI-powered robotic super soldier will never charge a machine gun nest to rescue an injured comrade because it simply doesn’t care or have any friends.
CHATGPT IS THE WORST “EMPLOYEE” EVER
So how does the previous rant about stripping the emotions out of battle apply to anything? After all, most of us (hopefully) are not asking our employees to charge machine-gun emplacements, so courage isn’t a factor. And some employers think they would love emotionless employees that complete tasks without complaints, asking questions or ever wanting to take time off to spend with friends or family. Navigating through employees’ emotions and keeping morale high and workers motivated can be difficult for even the most skilled managers.
Like with battle, removing emotions from the workplace can have large-scale consequences. Imagine employees that don’t care if you are satisfied with the quality of their work, have no genuine desire to improve, could care less about being creative and are never innovative. And they would never say they didn’t know the answer or how to do something. Instead, they would “create” an answer that might be complete nonsense and attempt to do any task even if they had no idea what they were doing. For most of us, that sounds like the worst employees ever, but that is precisely what ChatGPT currently is. Completely devoid of emotion, it executes tasks without the internal drive to achieve excellence that quality employees possess. Yes, the technology may improve, but relying heavily on ChatGPT or AI and expecting it to replace most of your marketing team is currently a recipe for mediocrity.
GOOGLE HAS ENTERED THE CHAT
Another wildcard in the whole ChatGPT mix is Google’s reaction to it. At first, the world’s largest search took a benign view of ChatGPT, saying it wouldn’t penalize content created by it. But that was before its popularity exploded, and people began to use it more extensively. Why go to Google to do research for a term paper when you can have ChatGPT write it instead? Or you could make the same “search inquiry” to ChatGPT and receive and reply in a more interactive and conversational manner. Gmail creator Paul Buchheit boldly predicted that ChatGPT and other AI chatbots would destroy Google and search engines in two years.
Maybe it just took a while for the giant to wake up, but it is clear now that Google sees the threat. Google’s hasty presentation of its own AI chatbot named “Bard” cost the company $170 billion in stock valuation when the inaccuracy of its answers called its competency into question. But Google doesn’t necessarily have to perfect its own chatbot or anti-AI AI to counter ChatGPT. Those that have been in the SEO trenches long enough will remember when authorship was a thing. The relationship between who wrote an article and how it ranks in SEO has never been made completely clear by Google. But it would make sense as E-A-T (expertise, authoritativeness and trustworthiness) are integral parts of Google’s Search Quality Rater Guidelines. After all, most of us would want to read an article written by a medical professional or doctor when searching for medical information rather than an AI or blogger who watched a couple of episodes of “House.”
As recently as 2020, Google filed a patent for Author Vectors, an advanced program that helps it identify who created content without an author by evaluating writing styles and levels of expertise. A simple tweak of Google’s algorithm could make this much more critical and suppress AI-generated content in its search rankings. Using ChatGPT to help with content marketing might be tempting, but it would be pointless if those AI-generated articles were buried in search results. And make no mistake; Google will probably do just that and heavily punish AI content.
THE IMPORTANT LESSONS OF 7/11 SUSHI
Many 7/11 mini-marts sell sushi. It kind of looks and tastes like sushi. It is not horrible, but it is not very good either and mainly exists for those desperate for a sushi fix. Sushi might seem as simple as chopping up some fish and wrapping it in seaweed with rice; that is what 7/11 sushi essentially is. But a lot of craftsmanship, subtleties, knowledge and artistry goes into creating genuinely great sushi. A perfect primer on the complexities of sushi is Netflix’s “Jiro Dreams of Sushi,” which profiles 85-year-old Jiro Ono. His Michelin three-star Sukiyabashi Jiro restaurant located in a Tokyo subway station commands a minimum of $300 a plate. Regarded as a master of sushi, Jiro proudly states that he is not perfect and that there is always room for improvement.
ChatGPT is currently the equivalent of 7/11 sushi. It can make content, graphics, digital art and other creations that kind of look like the real thing. What ChatGPT creates might be good enough for the desperate, but the reality is that it is not that great. Unlike sushi master Jiro Ono, ChatGPT doesn’t possess a strong internal drive to constantly improve its craft or achieve excellence. 7/11 sushi and ChatGPT might be convenient, but both are unsatisfying and are poor facsimiles. And integrating ChatGPT into marketing efforts will produce the same results as consuming too much 7/11 sushi.
THAT ELUSIVE HUMAN TORCH
The story of mini-mart sushi and the masterful creations of Jiro Ono also have parallels with marketing. Just like with sushi, there are different levels of marketing expertise and efforts. Whether ChatGPT can replace marketing teams depends entirely on what level they are operating on. Budget and minimal-effort marketers relying on low-quality strategies are definitely under threat because their work is easier to automate and replicate. ChatGPT might not be able to completely mimic their efforts yet, but it is only the first shot in the AI-powered bot salvo. The real threat is the next generation of improved versions that come after ChatGPT that could permanently eliminate low-level marketers and tactics.
High-level, engaging and effective marketing that actually delivers ROI, builds brands and sells products is a different story. Like the exquisite sushi made by Ono, it is complicated, intricate, challenging and not easy to replicate. There are many facets and strategies to marketing, but at its core, it is about creating a powerful connection between a brand and a customer. Doing it competently takes a great deal of skill, experience and deep insight into the human psyche. Right now, ChatGPT is far from being able to decipher that elusive combination. Will future AI-powered bots be able to? Predicting the evolution of technology is always a guess. However, we venture that high-quality marketing that delivers results will need a skilled human touch for the foreseeable future. Thankfully, most of us won’t need to switch careers soon.
WE ARE HUMAN AND HERE TO HELP
At Kahn Media, we keep up with the latest technologies and have our eyes on AI. But we also pride ourselves on not taking shortcuts or a cookie-cutter and automated approach to marketing. We have an entire staff of humans that are experts in digital marketing, social media management, influencer relations, public relations, event management, graphic design, photography, videography and content creation. As a fully-integrated agency, we create customized marketing solutions individualized to your brand’s needs that no AI bot can match. Contact us and speak to an actual human to see what we can do for your brand.