“AI Doesn’t Kill Jobs? Tell That to Freelancers: There’s Now Data to Back up What Freelancers Have Been Saying for Months”, 2024-06-21 (; backlinks; similar):
Jennifer Kelly, a freelance copywriter in the picturesque New England town of Walpole, New Hampshire feels bad for any young people who might try to follow in her footsteps. Not long after OpenAI’s ChatGPT made its debut, financial advisers who had depended on her 30 years of experience writing about wealth management stopped calling. New clients failed to replace them. Her income dried up almost completely.
When she asked, the clients she lost insisted they weren’t using artificial intelligence. But then, months later, some came back to her with an unusual request. The copy they’d been using AI to generate, they sheepishly admitted, wasn’t very good—and could she make it better? “It’s not a fix”, she says of the empty-headed, generic pabulum that AI [ie. ChatGPT & tuned LLMs] excels at writing. “You redo it.”
…Kelly, the copywriter in New Hampshire, is glad that at 62, she won’t have to endure many more years of being asked why she doesn’t use AI to speed up her work, or to clean up the dreck it generates. “We’ll be OK—our house is paid for, and I can get Social Security”, she says.
But the way that writing by humans is being replaced by what she sees as inferior material generated by AI still irks her. AI-generated content might still rank in Google search, but having seen so much of it, she can now spot it easily. “When I see something that looks like it was written by AI, I just switch off”, she adds. “The internet has just gotten so much duller.”
…Not long after ChatGPT debuted in November 2022, David Erik Nelson, a freelance sales and marketing copywriter in Ann Arbor, Michigan saw a jump in inquiries. “I was picking up new clients whose specific complaint was that their previous vendor had been giving them AI-generated content, and hadn’t been straightforward about it”, says Nelson. The AI had produced smooth prose intended for sales materials, but it was so generic, and often wrong, that it wasn’t about to convince people making 6–7 figure purchasing decisions.