If you’ve casually read some of my writings since generative AI gained popularity with the launch of smarter chatbots in 2022, you might get the mistaken idea that I’m a Luddite who doesn’t use generative AI, or worse, that I think it’s inherently “bad,” as one person told me after skimming a particularly passionate blog post about AI.
People of my vintage experiment with the latest technologies and platforms all the time. It’s a practice I cultivated when I was helping cover the explosion of the internet as a founding editor at the Industry Standard magazine, and now as someone who helps professional and nonprofessional authors write well. I do it because learning about new technologies makes me better at what I do and informs the content strategies I recommend to clients.
This current wave of gen AI is proving itself to be well along the path of Gartner’s hype cycle. “AI Has Been Massively Overhyped,” shouted a Wall Street Journal print headline from June 1, 2024, as one example. We’re now in the “trough of disillusionment” before gen AI matures, proves what it’s good at, figures out how it will make money, and enters the “plateau of productivity.” As with breakthrough technologies that have come before, there’s often some truth in the critique, and with enough time and experimentation, some truth in the hype. The devil lies in the details of these truths.
Personally, I don’t use gen AI to actually do the writing and editing of books, articles, and reports for clients. It still has some critical system errors for this journalist, including its well-known problems with getting things wrong, infringing on the intellectual property of others (including professional writers like myself), and being wide open to cybersecurity threats. (These aren’t my critiques—they are the top three risks holding back enterprise adoption for the more than 1,300 executives McKinsey surveyed in May 2024 who are experimenting with gen AI.)
In the early days of gen AI, it was fun to mess around with what are basically parlor tricks, like getting a chatbot to say crazy things, craft polite emails, recommend good restaurants, or plan a vacation. I’ve done all that and more, but I don’t rely on a chatbot’s output, mainly because the gen AI models I’ve played with so far have not proven to be absolutely reliable, too often don’t show their work well, need lots and lots of handholding to improve performance, and don’t always seem to understand me and my needs. Sure, it’s interesting to generate scads of random ideas, but I prefer to get mission-critical advice and results from someone I trust.
What I Found Gen AI to Be Good At
That said, there’s one use case in which gen AI excels—translation—and that capability has been a game-changer for my work. Perhaps reading my story will “generate” ideas about where new technologies might be a game-changer as you develop content and more generally do the knowledge work that powers much of the economy.
Writers and editors like me translate every day for a living. We translate complex ideas into smooth, accessible prose for a non-expert audience. Even when we work with native English speakers as ghostwriters and editors, many nonprofessional authors are simply unable to write in plain English for a wide audience. They therefore reach a much smaller number of people than they could if they wrote in a more accessible way. (Often that’s a feature of academic and technical writing, and not a bug.)
AI has been doing machine translation for a long time, and not always well, as users of popular translation apps may have experienced. But the newer language models naturally excel at translation, I’ve found in my writing and editing work. For technical reasons I won’t bore you with, translation is one of gen AI’s sweet spots, and one where specialist models are getting better at a rapid pace.
In this article, I’m talking specifically about two categories of translation I’ve experienced: foreign languages and dense specialist jargon.
Translating Foreign Languages
For translating literary works into other languages, where nuance and expertise matter a great deal, AI will never replace consummate artists like the Spanish poetry professor and translator I recently got to know. But on a quick first pass, it can speed up mundane translation tasks like understanding the gist of a webpage in another language, or translating customer service requests or user reviews for a travel or product website.
For a long-form article about my experiences in the former East Germany that I’ve spent years reporting and writing, I’ve had to conduct interviews in official German with government officials and other sources who don’t always speak English regularly. I’ve used ChatGPT-4 to translate emails, documents, and entire interview transcripts to and from English and German. Using older, non-gen-AI machine translation tools like Google Translate, built into the Chrome browser and available as an app, I’ve also read many more German news articles and academic books than I could have read given my plodding comprehension.
I studied German language and literature in college, and studied abroad on the only exchange program with East Germany before the Wall came down. Because I have kept my language skills alive through regular communication and visits with an old friend over the years, I still speak and read German in workmanlike fashion. But it’s very slow and painful for me to write in German.
It’s safe to say that I never could have accomplished the complicated magazine reporting project I recently completed without gen AI. The project required a level of writing ability to communicate effectively and quickly with sources that I just don’t have anymore. My sources didn’t seem to find fault with my writing, first in English and then in German using ChatGPT’s Translate GPT service. I’ve used it so much that I regularly exhaust the free daily usage limits.
Because I am at an intermediate level with the language, I knew the gen AI translation was mostly quite good and reliable, even in idiomatic German, unlike Google Translate’s often literal, wooden machine translations using neural machine translation (NMT) technology that launched in 2016. (For all the geeky details, Intento conducts a helpful report each year on the rapidly improving performance of models and engines. When PC Magazine tested AI chatbots vs. Google Translate with professional translators in 2024, AI chatbots won nearly every time—except in German.)
Translating Dense Specialist Jargon
In the second category of translation for specialist writing, I often translate dense technical, economic, and financial jargon into plain English, often based on interviews with subject matter experts at companies and academics at business schools. Over the last year I’ve often used gen AI to translate particularly dense passages that I’m only halfway comprehending after reading, and sometimes, rereading. I’m not an academic and it’s been 20 years since I got my MBA, after all. And yes, gen AI sometimes even helps me produce books, articles, and reports about gen AI for AI-focused companies.
I specifically give ChatGPT a “prompt” to “translate this article about [detailed subject] into plain English.” The results are strikingly accurate and easy to follow, even when I forget to tell the bot to act “as if” it were a particular type of professional, as people often recommend. I then go back and forth with the chatbot to further clarify and translate complicated jargon. It’s like having an on-demand tutor. I showed the results to one assigning editor, who was also struggling to understand a dense academic paper he’d assigned me to write about in a way that might get the attention of readers and the press, and he agreed it helped a lot.
Previously I would have had to spend a lot more time talking with subject matter experts to translate dense jargon and concepts, and would have gone through a few more iterations to make sure the facts and interpretations were absolutely correct. Iterations are still needed, but they have become a lot quicker when so much is already factually correct. To be clear, I don’t use the exact language gen AI provides in my writing, given the professional risks noted above about plagiarism, intellectual property, and accuracy. But along with a lot of other research, traditional search queries, and interviews with experts, these gen AI translations inform my writing and allow me to do what I do much faster than using traditional methods alone.
The Future of Work
Will gen AI replace me? At first I was worried that that might be true, and when work began to dry up for me as people played around with AI for writing and editing, I got really worried. Now that many organizations have tried and failed to produce high-level thought leadership with gen AI, they are sometimes coming back to us humans for help.
As I’ve long written, I don’t think I have too much to worry about over the near-term. (Long-term is anyone’s guess, and I might be retired by then.) Gen AI’s often bland, formulaic language and lack of creative thinking abilities is not a replacement for the sophisticated narratives I create based on decades of experience as a journalist and editor. But it does augment me, as I helped authors argue in a recent AI book about machine-human collaboration.
The key in both these use cases is that I already speak the foreign and technical languages I’m using this particular tech tool to help translate. I have enough mastery of language, business, and technology from many years of experience to know when something is good. When it’s good, the results can speed up my work by multiple factors. When it’s bad, at least it didn’t take long to find out.
AI is not a replacement for my skills—at pushing authors to think hard and write well and therefore produce measurable business results—but rather it is a force multiplier of my unique human perspective and experience.
Of course, gen AI is getting better at lots of things every day. So I’m genuinely curious how you are using new technologies like AI to innovate, open up new growth opportunities, and amplify your unique human talents. Because the devil is in the individual details for each of us.
Mickey Butts collaborates with smart people at the top of their game to ghostwrite and edit great narratives that land with the C-suite and foster trust that genuinely opens doors. In the process, he helps launch platforms that generate speaking, consulting, press, and sales opportunities that produce substantial ROI. If you’d like to discuss the ideas in this article or Mickey’s work at the Insight Content Lab, please feel free to be in touch at mickey@insightcontentlab.com.
