Writers: If A.I. can’t do your job, watch out.
We need to give A.I. our chores so we can do our jobs.
I fed my rough draft of today’s article into Bard, Google’s artificial intelligence chatbot, and asked it for edits and proposed headlines.
Within seconds, I had a list of suggestions to squeeze my article into a generic formula. The vapid headlines hardly described my content.
Experiments like this have comforted too many writers since Chat GPT became a household name last winter. We’ve snuck over to the chatbot without telling anyone, given it a writing assignment, and breathed a sigh of relief at the results. Good news: Generative A.I. might sound human, but it doesn’t sound like a great writer.
True, the technology has plenty of shortcomings. Its imagination is limited — except when asked for facts that it doesn’t know, in which case its imagination clearly comes up with something. Much of its prose is pathetic and formulaic, and its poetry is just. Plain. Bad. There’s also the ethical question — isn’t the A.I. just stealing someone else’s work and repackaging it?
That sounds like job security for me. In the words of Brian Regan, “I’d like to yield the remainder of my time to my opponent.”
But now is no time to let our guard down. The way writers and other creatives choose to embrace, ignore, utilize or criticize generative A.I. today will determine how we fare in a few years when A.I. has improved exponentially.
A.I. and patterns of disruption
The common writer’s reaction to generative A.I. reminds me of patterns that the Harvard Business scholar Clayton Christensen noticed when he studied the 1980s disk drive industry. His book The Innovator’s Delimma describes1 this cycle:
Company A makes a large hard disk with 200 MB of storage, and it’s huge.
Company B starts making a smaller drive that only holds 100 MB.
Company A ignores it because its customers need 200 MB hard drives.
Company B improves its hardware until the smaller disk can hold 200 MB.
Company A loses its customers.
Company C makes an even smaller hard drive that only stores 50 MB.
Company B ignores Company C’s innovation ….
You can probably see where that’s going. Today, a sliver-thin device smaller than my pinkie nail holds more data than the hard drive on the Tandy 1000 I used for Tetris 30 years ago. Companies A and B are out of business.
The dilemma Christensen described was this: Satisfying your customers’ immediate needs may prevent you from innovating, which almost always involves creating something that seems inferior at first. Ignoring an inferior competitor leaves you stranded in the status quo.
Eventually, the tech wizards in Silicon Valley will take care of A.I.’s “hallucination” problem, which is when bots make up facts from whole cloth. They’ll also find ways to make A.I. more creative and more enjoyable to read. When that happens, those who are rolling their eyes today will find Chat GPT returning the gesture.
If we understand the job that we creatives are called to do and then approach A.I. for innovation, rather than reassurance, then A.I. will take our chores, but not our jobs.
How writers survive the A.I. revolution
Understanding the job to be done
Instead of scoffing at A.I., we need to find out where A.I. needs us. Or, more specifically, where does a world that has A.I. need us?
Another book by Clayton Christensen, The Innovator’s Solution, described what companies could do to break the cycle of disruption, or at least get on the disrupter side. One of the core principles is to understand the “job to be done.”
Consider his example of milkshakes. If you ask customers what they like in a milkshake, you’ll get competing answers. Some want it flowing but with chunks of fruit and chocolate, others want it thick and smooth. Averaging these wishes gives you a mediocre milkshake fit for the dumpster.
But if you instead ask customers when and why they buy a milkshake, you learn what “job” they are “hiring” the milkshake to do. Parents who are buying a milkshake for their kid have different needs than an adult drinking the milkshake for breakfast on the drive to work. Understanding this “job” allows you to design a menu that gets the job done.
Likewise, those of us who wield words need to understand the job we’re hired to do. If we see our job as solely to write content, then we might only innovate within that window. Work would become a contest of stringing words together beautifully, engineering the words to get the idea across succinctly, and telling a compelling story. And A.I. is coming for us, because will learning to do all that, fast. Ignoring A.I. will leave us stranded in the status quo.
What is a writer’s job, really?
What, then, is our job? What do people actually need us to do? Ultimately, our job is to provide content that meets the needs of the information consumer — the reader, watcher and listener. Within this mission, there are many smaller jobs.
Here are a few “jobs” that writers fulfill:
My job is to discover information and provide it to my readers. I’m an editor and curator.
My job is to help people sort through the noise and find meaningful, reliable information. I inspire trust and create meaning.
My job is to inspire people to take action that helps them reach their goals; articles are just a vehicle for that.
(Since I work in higher education marketing) My job is to help students discover the academic program that is right for them.
My job is to think through competing ideas, find the merits in each, and select one to champion in my next article.
Understanding these jobs to be done allows us to focus on the tasks that ChatGPT would struggle with. They’re more difficult than writing a formulaic blog post. They’re also more inspiring for us humans.
Whitney Johnson, a Christensen disciple who applies disruption principles to personal growth, would call this “playing to your distinctive strengths,” part of her seven point framework described in Disrupt Yourself.
Letting A.I. take our chores
Understanding our job helps us find out what A.I.’s job is — how machine algorithms can help us fulfill our job.
Earlier I mentioned that I was unimpressed with Bard’s suggestions for this article. But one day last week, I took pasted in a long email a professor had written to me and asked Bard to turn it into a 400-word story. I told Bard what the focus should be, and I said the first paragraph should be something along the lines of X.
Within a few seconds, I had a draft.
I asked for a couple of edits, including asking Bard to select a couple of direct quotes from the original material to include in the story. It took me fewer than 15 minutes to get to a good first draft of that article ready for my editing.
In this case, A.I. took my chore of writing and rewriting. I got to keep the job of deciding what was meaningful and important for my audience.
If we understand the job that we creatives are called to do and then approach A.I. for innovation, rather than reassurance, then A.I. will take our chores, but not our jobs. But if we ignore A.I. except to laugh at its weaknesses and wring our hands over its strengths, our jobs are in peril, too.
Further reading
Michael Smart describes how public relations professionals can outsource work to Chat GPT. I had written most of this article prior to reading it, but he inspired my idea about letting A.I. take our chores.
Tim Woods shares the 8 principles of The Innovator’s Solution.
Read Christensen’s books yourself: The Innovator’s Dilemma and The Innovator’s Solution.
Check out Whitney Johnson’s books including Disrupt Yourself and Smart Growth.
All numbers in my description of the hard drive industry were made up.
Thanks for this article with the great resources you used. I hope to read some of them to understand “disruption”.