I love Seth Godin. I admire and respect him. I read his blog every day. And, by the way, he is just as nice in person as he is on screen. And his advice is usually 100% on point.
But his blog on August 21 gave me a perfect example of why we’re not using AI to write.
And before I tell this story, I should say that his blog article was simply saying that if you have to write as part of your job, and you’re not a writer by trade, AI (specifically, Claude.ai) can help you.
Seth recommends:
If the writing you’re doing doesn’t need to be in an idiosyncratic voice, take your memo, paste it into claude.ai and say, “please rewrite this to make it clear, cogent, positive and concise.”
Note he did say that “if the writing you’re doing doesn’t need to be in an idiosyncratic voice,” which definitely gives him a pass on this one.
But his blog gave me an idea.
I’m writing a new book (Love, Leverage, and Logistics: How You and Your Business Can Change the World), and yep, it’s definitely idiosyncratic (which comes from the Greek, meaning “one’s own”).
After reading Seth’s blog entry, I decided to try a little experiment with this ai copy generator. I took a small section of the book where I discuss the damage that jerks do to their companies. The subhead above this copy is “Being the boss doesn’t give you license to be a jerk.”
Jerks who start companies treat their employees, customers, and partners in unkind ways. They treat these people like they don’t have a choice.
But, more than ever, thanks to our digital world, people have plenty of choices. People can easily find another job. Customers can find alternative products and services. Partners can discover other companies they can represent. All of these opportunities are just a click away.
Only the most insecure people, or people with certain circumstances, will continue to work for a jerk.
Insecure people will never talk back to the boss, and will put up with all kinds of insulting or unethical behavior. They will never tell the boss about problems that come up or they will minimize them. The boss will never get the whole picture. His or her decisions will be misinformed, erroneous, and could even be downright dangerous.
People with certain circumstances—such as needing to remain in a particular geographic area to help an aging parent—will change jobs as soon as those circumstances change.
I did exactly as Seth recommended, asking Claude.ai to “please rewrite this to make it clear, cogent, positive and concise.”
Here’s what Claude.ai came back with:
Leaders who treat employees, customers, and partners with respect foster better relationships and business outcomes. In today’s digital world, people have more choices than ever for employment, products, and partnerships.
Companies that prioritize ethical behavior and positive work environments attract and retain talented, confident employees. These team members are more likely to communicate openly, provide valuable feedback, and contribute to well-informed decision-making.
While some individuals may temporarily stay in suboptimal work situations due to personal circumstances, most seek opportunities that align with their values and professional goals. Ultimately, businesses that cultivate a culture of mutual respect and open communication are better positioned for long-term success.
This is what I observed by the changes that Claude made.
1. The word “jerks” was completely removed. I believe I can safely assume that is a trigger word, a no-no word, a word that signifies that someone is calling someone else a jerk.
My intent in writing this book is to encourage and inspire the kind, non-jerk people in the world who have started their own businesses and want to create a culture of positivity and caring.
The world really is divided between jerks (who make life harder for everyone else) and kind people (who strive to make things easier for others).
Pointing that out helps kind people realize that they are not the problem when they’re working for a jerk. Taking the word “jerk” out of this copy removes this entire concept.
2. The content lost its human-to-human aspect. This is the main difficulty I have with AI-generated content. It reminds me of content written by people in one country aimed at readers who live in another country.
IBM once hired me to evaluate content written by people in one “offshore” country aimed at readers in America. I spent hours going through massive amounts of content.
It was painful.
The problem wasn’t the talents of the offshore writers. The problem was that their culture and their own way of speaking English was built into their writing.
I’m not a writing snob. The truth is, the opposite is also true. If someone who grew up in the U.S. was hired to write copy aimed at people in the offshore country, those readers would know that someone from the U.S. had written that copy, just as the U.S. readers knew that the copy was written by an offshore country. They’d even be able to guess which country it was.
AI bots are not humans. They can do amazing things; the images they generate when properly prompted blow my mind. (The only problem there is their difficulty with appendages—arms, legs, fingers, and toes. I’m sure they will get this right eventually.) But they can’t feel what we feel, with our souls and hearts. They can’t empathize. They can pretend to empathize, but it’s a false display of intuition; more like they are simply trained to appear polite and interested.
3. The content is bland and vanilla. When I first started writing articles about marketing, my husband, who edited everything I wrote, kept pointing out something he called “marketing speak.” It was passive, lifeless, and very third-party-ish. He drilled that out of me, and I’ve never looked back.
Copy we write for our clients isn’t quite as “idiosyncratic” as my book copy, but it definitely is not “marketing speak.” It’s personal and relevant to the reader. It does reflect the beliefs and experience of our clients, whom we interview about that particular blog topic before we write. And because we interview their customers before we write any copy, we know who we are writing to and can write more conversationally and personally.
What Claude came up with could have been found in ANY book on management. It’s nothing new; it’s nothing you could take back to your office and put to work; it’s just words strung together without those insights that come from personal experience. It’s a book report.
How do I feel about AI overall? The way I’ve felt about every tech advancement since I started making tech a major part of my life years ago.
It’s a tool; when the tool is used for what it does best, by people who want to make things easier for other people, it’s fine.
When it is touted as the answer to all the world’s problems, or misused by jerks, it’s definitely not fine.
