What happens after human storytellers help chatbots seem more human?
Proponents of AI repeatedly say how the rapidly advancing tech will create new jobs. They’re not wrong, but the type of AI jobs for traditional writers may not be what you’d expect.
Many of the AI jobs you’ll find on career platforms like LinkedIn are for “AI writers.” When I searched “AI” among Canadian jobs on LinkedIn today, I got roughly 8,000 results including several writing roles.
That sounds promising to those of us who have decades of experience as a writer/editor, and are now learning how to navigate Large Language Models (LLMs) to enhance or even rate their own content.
A new definition of ‘writer’ for the modern age?
However, in order to understand exactly what an “AI writer” is according to employers, you have to look at the fine print of each job posting. Some of the employers are asking for skills beyond what you’d expect from a writer, such as analyzing data.
Some of these roles don’t really involve writing at all, but rather training AI chatbots to tech them how to respond better to human queries. Most of the other jobs ranking high in the search are similar: they use the word “writer” in the job title, but it’s clear you’ll just be having conversations with AI to train their models.
You could argue AI “writing” positions now lean more towards editing AI outputs, as myself and an industry colleague previously predicted. Many of the advertised roles involve revising AI responses, as well as ranking how effective they are based on training parameters.
I’m not totally against taking on one of these “writing” roles, which are almost exclusively freelance/contract with no benefits. After all, AI writing models (which can be customized for specific audiences) are learning from humans one way or another, so I might as well get paid for doing it.
The pay range varies greatly among these AI “writing” roles — some offer as low as $13/hr for projects, while others offer $25/hr U.S. or more if you have a PhD or are willing to work in a hybrid model.
However, while I have roughly 20 years of experience in writing and editing for digital platforms, I’m apparently not qualified for several of these AI “writing” jobs.
That’s because many of them are asking for a bachelor’s degree as a minimum, whereas I “only” have a college diploma in journalism.
But what if I do land one of these AI “writing” jobs, which are actually AI training jobs? At what point will I train a model to be efficient enough to no longer need human input to improve?
It’s sort of like when your boss makes you train a new employee, only to replace you with the newbie (likely at a lower wage) when they’ve become proficient.
The AI is like the new guy in the office, soaking up our knowledge. If an employer deems they can get reasonable quality for less money, they won’t hesitate to axe their human capital.
Retrain yourself while training AI models
Yes, industries evolve. The onus is on workers to stay relevant in a changing work landscape. However, for traditional writers, the AI industry might be moving faster than many of us can keep up with.
The fact that I have decades of writing/editing expertise, and have been experimenting with AI tools for more than two years seems to be irrelevant to many employers in this age. If my CV is getting ignored by them, then it can’t be easy for those with less relevant experience than me.
But make no mistake: the better that AI chatbots/language generators get, the less need they will have for human trainers and editors. The human writing teams for AI companies will continue to shrink.
There are a few solutions for human writers: you can write original news articles, which will probably require you to have a flexible lifestyle. I can tell you as someone who was a reporter for several years that it eats up a lot of time and energy.
You could write novels, although your chances of wooing a big publisher as an unknown are fairly slim. You can share your human perspectives through blogging, although rampant AI might be devaluing this content.
I’m not trying to tell you to stop writing. We need humans to continue occupying this space. But be aware that future “writing” jobs may not be writing at all — and if you’re not prepared, someone with relatively little experience could scoop a job from right under you.