There are so many people weighing in on the ChatGPT AI topic that I’m going to skip all of the “this is a life-changing technology that could ruin the world (and take everyone’s job)” chatter. And, “It could be really bad if it falls into the hands of a Bad Guy” (paraphrasing Bill Gates).
Instead, I want to look at the practical aspects. How useful is this tool, really, for life and business use? How should we see it as marketers?
I decided to take the subjects I know a lot about and see what we can learn about the answers, comparing the AI answer with extensive experience. And yes, one of the three subjects is marketing. Of course.
I’m going to start with my first subject, though, just because it’s the first thing I asked ChatGPT and I was super curious about what it would say. I asked, “How do you sail upwind?” I wanted to compare the answer with what I’ve learned over decades of bay and ocean sailing, from windsurfing in San Francisco Bay to the ocean sailing we did on the 48-foot catamaran my husband and I had built in South Africa and then sailed home to New England.
This was the answer. No need to do more than skim it if you are not a sailor, but interesting to read it carefully if you are.
OK. First, we are missing the big picture. As any good teacher will tell you (and remember, the best marketing teaches rather than sells), you need to provide context before providing details. Human brains absorb content so much better if they have a “bucket” to put it in.
So the big picture that the chatbot fails to mention is this: You can’t steer a boat unless there is water flowing past the rudder. It’s not like a car, where you can actually turn the wheels while standing still. On a sailboat, if you are not moving along, and you try to turn the rudder, nothing will happen.
Well, that’s not exactly true. What will happen is you will lose control of the boat. You will be at the mercy of the wind and the current, and if you are close to land (which you usually are when you tack, because you need to turn the boat before you hit the land), you will end up on the rocks.
They also fail to mention that, to make sure you actually make it through the turn, you should get going a little faster before you start to make the turn. So you turn away from the wind just slightly, get going a little faster, and then make the turn.
And, this is a smaller point, but they also say: “During the tack, the jib/genoa (that’s the smaller sail in front of the mast) will need to be released, passed over the forestay, and then trimmed in on the new side.” Once again, if you could follow this advice, you would be in a Big Dangerous Mess.
The forestay is the wire that goes from the top of the mast to the bow. The leading edge of the jib is attached to the forestay.
In fact, it’s impossible to have a sail pass in front of the forestay when the wind is coming from the front of the boat, because the wind is pushing the sail back over the boat, behind the forestay.
So I guess you’re safe in this case, because you can’t actually do what the bot tells you to do, when sailing into the wind. If you were sailing downwind, with the wind coming from the back of the boat, sure, you could have the jib try to pass in front of the forestay. But that would be another Big Mess.
The main point?
In this short segment of copy, a chatbot has just given you three bits of dangerous or impossible advice. Dangerous because if you end up on the rocks with the waves and wind banging your boat into the rocks, you will wreck the boat and you could die.
Now, of course, this is a learning app, so I checked a few weeks later and it had revised its answer:
Note that this time, there was no mention of tacking. Or advice about tacking. So while there were some hints about speeding up and slowing down, there was nothing about turning the boat when you get close to land as you continue to sail upwind. Obviously, when you get close to land, you have to turn the boat. Still a problematic answer.
Based on these answers, at this point in the life of ChatGPT AI, if you are taking its advice—and only its advice—you have decided to live dangerously.
Personally, I’d give ChatGPT an F on this one.
OK. Next subject: Marketing advice from ChatGPT.
Simple question: How do I market my product?
This is Marketing 101, a solid, to-the-point book report on the steps involved in marketing. The only thing wrong with this is what is wrong with all marketing: If you never interview your customers to find out what they were looking for when they set out to buy and why they bought from you, everything you do will miss the mark, and not pay off for you.
You can do all the steps above, but when your potential customers come to your site or see your product for sale on another site, they won’t “see themselves,” and they will click away.
You need to know what their Mindset is when they set out to buy, which I define as their specific desires, concerns, and questions.
Guessing what these are is the number one cause of marketing failure, in my experience.
So, once again. If you take this advice—and only this advice—you have decided to live dangerously.
I give this one a C-, because at least you won’t die following this advice. But you could still end up in a big mess, with more money going out than coming in because your marketing wasn’t working.
Another question: How to stay happily married.
It took us a number of years to end up with a happy-every-single-day marriage, but we did pull it off, even with fatal cancer and other life challenges thrown in. We were more in love at the end than we were at the beginning, and the beginning was pretty awesome.
So, once again, I wanted to see what ChatGPT had to say and compare it to my personal experience. Surprisingly, even though the chatbot apologized for not having “personal experiences or emotions,” the answer was pretty darn good.
Not bad. The only thing it leaves out is you have to be careful who you marry in the first place. If you marry a jerk or a diva, none of this great advice will work. Jerks and divas enjoy making life harder for everyone else, so no matter how much you do all the steps listed above, it will never be enough.
I’ll give this one a B.
How ChatGPT handles buyer questions
Personally, I think Google has more to lose than anyone with the release of ChatGPT. Google is not very good at providing agenda-free answers to “what is the best [whatever].” Google makes about 80% of its income through advertising, so this is going to be very disruptive if everyone turns to AI tools to get the answers they would have gone to Google to get.
With this in mind, I wanted to see how AI handles typical buyer questions.
I drive a Mazda convertible because it was the best-handling convertible we drove when looking for a new ragtop. And, I thought ChatGPT gave a great answer to this question, based on our test drives:
I’d give this answer a B+ or even an A-.
On the other hand, when I was looking for a replacement charger for an electric dinghy motor made in France (“Temo”), and kept coming up empty on Google, I checked in with ChatGPT, and was just as disappointed. First it had to ask for more information about the Temo (which I don’t think is a minus; best that it understands what I’m looking for before answering), and after I provided that, it told me to do what I had already done:
I guess this is a D. At least the advice it gave was common sense and didn’t leave anything out. But it certainly didn’t find anything helpful.
Conclusions about ChatGPT AI
1. This is the fastest-moving tech developed so far. Our own team is learning all it and experimenting as fast as we can, to make sure we are using AI to best advantage for our clients.
2. We are in the first phase of the technology cycle on this one, where the hype outpaces the reality. As people start using it more and learning what it can and cannot do, realities will start to outpace the hype.
3. AI will definitely disrupt people, companies, and whole industries, just as the web—and Google—did. The trick is to learn fast, observe what new needs are created, and figure out where you could match your learning with those needs.
In other words, be of service to others. This is always where the real money is.
In any case, it is going to be a wild ride.
From all that I’ve seen or read about ChatGPT, beside missing, incomplete, or inaccurate content that inquiries can produce, the related concerns are the time and effort involved to have material or guidance which is truly valid and useful. It’s a bit like trying to fact-check a politician giving a barrage of statements during an interview. The reality, falsity, and in-between of what is communicated can’t be done in real-time, even with a significant investment of resources (people, science, resources, money, etc.). It requires back-end analytical and time commitment, often not available for marketers. If, at the end of the day, a grade of B, C-, or F is the best that ChatGPT can produce, and not reliably and consistently an A result, where’s the net value?
Excellent post! It seems ChatGPT and GPT-4 fascinates us because it it so human-like. It isn’t human, of course. Computers don’t understand meaning; they can only find and repeat common patterns. Evaluating whether those patterns are actually meaningful (for the foreseeable future, anyway) must fall to us.
This is a terrific read. The examples are spot on. We need more insightful post like this on this controversial topic. My concern is less about ChatGPT’s capacity in general and more about how it will impact customer experience. As I wrote in an article earlier this month for Forbes.com, can a ChatGPT and its AI cousins yet to be created, generate a customer experience that leaves you with such a profound and compelling memory you will want to encourage your friends to join you? Here are four features ChatGPT/AI cannot do.
ChatGPT/AI Cannot Serve with Generosity.
As captivating as the Wall-E and ET characters were to moviegoers, AI has no heart. According to researchers, even after significant advances, AI is still far from being able to feel emotions in the same way humans do. “They can recognize and respond to emotions, but they do not have the same complex psychological and physiological responses humans do,” says software engineer Bimo Tristianto in MLearning.ai.
ChatGPT/AI Cannot Anticipate Customer Expectations.
Future perfect thinking requires a deep understanding of customers and an intuitive leap of faith in what they want or need. The two astronauts in the 1968 hit movie 2001: A Space Odyssey defeated a rogue computer named Hal, not by doing what they had been trained to do, but through an innovative solution that defied conventional machine logic. According to research, ChatGPT cannot intuit. Its responses vary widely in reaction to tiny differences in how questions are phrased.
ChatGPT/AI Cannot Forgive Customers
“The customer is not always right,” said Stew Leonard, Jr., CEO of the famous grocery chain that bears his father’s name. “But it’s our job to make them feel right.” Acceptance and forgiveness are vital attributes in the customer loyalty world. Laura DeCarlo, of Career Directors International wrote, “When things go wrong, assume the client simply wasn’t paying attention and be prepared to guide them with kindness, compassion, and professionalism.” ChatGPT/AI is wired to be accurate but not empathetic. It has no soul.
ChatGPT/AI Cannot Demonstrate Authentic Respect.
Respect is more than kindness; it is a demonstration of valuing–actions that communicate, “I consider you to be important.” The machine cannot put itself in the position of the person it is judging. “It is the inability not only to empathize but then employ that empathy to generate additional reasons to act (or not act) that makes the machine’s judgment impersonal and pre-determined,” wrote Eliav Lieblich and Eyal Benvenisti.
AI and ChatGPT’s place in customer service support will be important. However, never forget what the “A” in AI stands for. Since it cannot provide the same kind of personalization as a human, it will live more successfully in the “backroom” with coders, researchers, and auditors, leaving the front of the house for people who can genuinely serve customers with heart, soul, anticipation, and respect.
Portions of this article were previously published as “Can AI Deliver the Kind of Customer Experience a Grandmother Would” by Chip R. Bell, Forbes.com, May 13, 2023. https://tinyurl.com/39t2fcwn
Great piece, it highlights the limitations of the tool very well. The AI bubble is already sagging, if not bursting. It can of course be very useful, but only with the right (human) direction, and only for certain things. I am reminded of a story – probably apocryphal – from the 1920s, in the early days of electrification (also touted as a great jobs destroyer…). Keen to get local businesses connected to the nascent grid, and earn commission in the process, an aspiring entrepreneur put up a sign that read, ‘Why kill yourself with work – let electricity do it!’
Just replace ‘electricity’ with ‘AI’ of course, and that’s the message today. Plus ça change…
CHAT GPT Certainly is an excellent Guide and Fast Knowledge Base or Library, but it can never substitute your experience and your personal touch in a speech or message you would like to read. For me, when I listen to a webinar, where the speaker is perfect, doesn’t pause, everything is perfect, it counter effects him, I feel everything is prepared, based, prepractice, and doesn’t come from a good vibe or good feeling.
Just my very humble opinion!!!
A tool built by human, threatening human intelligence to replace at work. In future many more such inventions are definitely likely to come. Its upto the people who deploy them be cautious. its like inventing a table for not dieing
Thanks, all, wonderful insights. They all make me think that while it may not be perfect—in the ways you all describe—it may be “good enough,” and we all know that “good enough” has dominated many markets for many years. It will be interesting where AI really ends up providing real value to society.