Generative AI and Thought Followship
When you have a machine chase what others have already written, you are no longer offering thought leadership in your marketing efforts.
The other day, I had put a query on one of the sites journalists frequently use to find potential new sources. It may be that you don't want to keep using the same people or that you're looking for a specific background match. The site in this case was Qwoted and I received a couple of responses that reminded me of what some other sites, like Help a Reporter, or HARO, have become: a place where all sorts of people with no particular expertise in a topic hold forth with material they’ve skimmed from the Internet as though that should be enough to earn them publicity.
Additionally, the HARO responses typically come in a form mechanical in structure and manner, as though stamped out for them by commercial software. The senders hope for a mention in a published story with their typical demand for a link back to their site as a form of SEO. The more connections back, the better their search rankings. Why else would someone who has a site offering information on tech products decide that they know the warehouse market, or another person with advice on “pool installation, landscaping and overall home maintenance” try a hand at listing important cybersecurity conferences? (Both real examples.)
To find this creeping into another site was disheartening, as I dislike the potential of losing a useful tool. But in this case, one of the responses was so empty and anodyne that I thought it could be AI generated. I asked the sender, who said it was not. The form bothered me still and I checked with several generative AI detectors, getting two out of three saying the material was a result of human creation, but still had no intention of using it. I just didn’t trust it.
Shortly after, I received a note from Qwoted that the “expert” had been taken off the system because their internal checks had determined that the material was from a generative AI program. I was pleased that they checked and take an active approach.
But all this started me thinking about what people try to achieve when they respond to a reporter’s request. The term these days in content marketing is “thought leadership.” Companies and people typically want to prove their expertise, and I know beyond much question that many are already using chat bots to create pieces that they can run on blogs, offer as white papers, or pass off as responses to the press. The people doing so are likely smug that they’re helping their goals and saving themselves time or the need to hire a writer who has expertise in such work.
They may indeed save money and time in the short run, but what they produce isn’t thought leadership. Instead, it is what I’m now calling thought followship. Generative AI delivers written results works by analyzing immense bodies of already existing writing and developing complex statistical models of how words follow one another for given contexts. That is why you could, as Manchester Institute for Education digital sociologist Mark Carrigan did, ask for a poem about generative AI and its impact on universities in the style of Dr. Seuss. Ironically, the language is too complex and doesn’t really sound like Dr. Seuss, but what do you expect for nothing?
And that’s the point. What does someone using this in a serious way to create “thought leadership” expect when all the writing, all the thinking, all the intellectual and emotional connections have been made before?
Some might argue that there is no original thought under the sun and that everything has been considered before. They would be incorrect. There are always people who have thoughts others haven’t previously conceived, at least not in that specific way. Socrates, Leonardo da Vinci, Hildegard of Bingen, J.S. Bach, Immanuel Kant, bell hooks, Alan Turing, Eugene Ionesco, Marie Curie, Shakespeare, Georgia O'Keeffe, Billie Holiday, Thelonious Monk — are all of them really nothing more than subtly adjusted copies of what went before? That assumption may make many comfortable with their own worldly status, but please.
Even outside such illustrious company, there are many people with insights into the arts, business, crafts, music, and life because they applied their long personal experience to some topic and developed the individual language and intellectual structure that allows them to state things in ways others haven’t or can’t.
If you aren’t a sui generis genius, you can at least aspire to strong and sometimes profound insights into what you know and live. Those are the vineyards of thought leadership.
To instruct a machine to create something that doesn’t come from your own self means the results literally come from following lines of words that others have put down. It is thought followship because you chase what others have done. This won’t provide the individuality and uniqueness that might draw some attention or actually help others who might have been waiting for decades for the lesson one specific person could impart in a way the reader finally grasped.
That consideration is in addition to many other issues, including the potential for copyright infringement, the full indemnification that the generative AI vendors demand so that if there is trouble, it is the one thing guaranteed to be yours. And the lack of copyright protection you have in the output, so anyone could come along and lift it for their own uses. Because, after all, no original thoughts and all that.