If you’re worried about how AI will affect your job, the world of copywriters may offer a glimpse of the future.
Writer Benjamin Miller – not his real name – was thriving in early 2023. He led a team of more than 60 writers and editors, publishing blog posts and articles to promote a tech company that packages and resells data on everything from real estate to used cars. “It was really engaging work,” Miller says, a chance to flex his creativity and collaborate with experts on a variety of subjects. But one day, Miller’s manager told him about a new project. “They wanted to use AI to cut down on costs,” he says. (Miller signed a non-disclosure agreement, and asked the BBC to withhold his and the company’s name.)
A month later, the business introduced an automated system. Miller’s manager would plug a headline for an article into an online form, an AI model would generate an outline based on that title, and Miller would get an alert on his computer. Instead of coming up with their own ideas, his writers would create articles around those outlines, and Miller would do a final edit before the stories were published. Miller only had a few months to adapt before he got news of a second layer of automation. Going forward, ChatGPT would write the articles in their entirety, and most of his team was fired. The few people remaining were left with an even less creative task: editing ChatGPT’s subpar text to make it sound more human.
By 2024, the company laid off the rest of Miller’s team, and he was alone. “All of a sudden I was just doing everyone’s job,” Miller says. Every day, he’d open the AI-written documents to fix the robot’s formulaic mistakes, churning out the work that used to employ dozens of people.
While I agree with most points you make, I cannot see a machine that is, at a bare minimum, able to translate between arbitrary languages become irrelevant anytime in the foreseeable future.
OK, I agree, translation is useful and is fundamentally something it makes sense for.
Disagree about “arbitrary”, you need a huge enough dataset for every language, and it’s not going to be that much better than 90’s machine translators though.
It’s closer, but in practice still requires a human to check the whole text. Which raises the question, why use it at all instead of a machine translator with more modest requirements.
And also this may poison smaller languages with translation artifacts becoming norm. Calque is one thing, here one can expect stuff of the “medieval monks mixing up Armorica and Armenia” kind (I fucking hate those of Armenians still perpetuating that single known mistake), only better masqueraded.