You can also Engineer AI Buy this book now

There is a hilarious, almost poetic irony unfolding in the tech world right now.

For the past few years, the internet has been flooded with doom and gloom predictions about how artificial intelligence is going to completely replace human writers, journalists, and copywriters. Yet, behind the scenes, OpenAI and other tech giants are spending millions of dollars to do the exact opposite: they are aggressively hiring thousands of human writers and paying them healthy hourly wages.

If ChatGPT is supposedly so smart, why does it need to hire humans to write for it?

The answer reveals a fascinating bottleneck in the AI race and it might just be the best news human creators have heard in years.

The “Data Drought” and the Wall of Text

To understand why OpenAI is hiring writers, you have to understand how ChatGPT was built in the first place.

Early AI models were trained by scraping the open internet books, blogs, Wikipedia pages, and Reddit threads. The AI read billions of pages written by humans and learned how to mimic our speech patterns.

But tech companies have run into two massive problems:

1. ​The internet is running out of good data. AI companies have scraped almost all the high-quality, publicly available English text on Earth.
2. ​The internet is full of garbage. Mimicking a random Reddit thread or a poorly written blog post makes the AI sound, well, generic and unreliable.

Worse yet, as the internet becomes flooded with AI generated content, future AI models risk being trained on their own output, leading to a digital degradation known as “model collapse.” To make ChatGPT smarter, more nuanced, and truly helpful, OpenAI doesn’t just need more text—it needs impeccable, verified, deeply human text.

What Do These AI Writers Actually Do?

If you get hired to write for an AI company, you aren’t writing standard articles or blog posts. You are participating in something called RLHF (Reinforcement Learning from Human Feedback) and Data Annotation.

Essentially, you are acting as a personal tutor to the algorithm. The work generally falls into three buckets:

• ​Writing the “Golden Standard”: A writer might be given a prompt like, “Explain the economic concept of inflation to a 10-year-old using a metaphor about arcade tickets.” The writer crafts a brilliant, creative, engaging response. This response is fed directly into the AI as the textbook example of a “perfect” answer.
• Fact Checking and Editing: The AI generates a draft, and a human expert reviews it line by line, fixing subtle logical errors, tone issues, or outright lies (hallucinations) that a machine can’t spot.
• ​Ranking Responses: The AI generates five different answers to the same prompt, and the human writer ranks them from best to worst, explaining why one answer feels warmer, clearer, or more accurate than another.

OpenAI isn’t just hiring creative writers, either. They are hiring poets to teach the AI emotional depth, software engineers to explain the logic behind lines of code, and lawyers and doctors to ensure the AI’s industry-specific advice doesn’t get dangerous.

The Ultimate Reality Check

​This mass-hiring spree proves a vital truth that the tech evangelists tried to ignore: Human creativity and intuition cannot be automated out of thin air.

​An AI cannot understand what makes a story move a reader to tears, or what makes an explanation suddenly click in a student’s mind. It can only look at the data we give it. Without human writers constantly feeding the machine high-quality original thought, creative metaphors, and rigorous fact-checking, the AI hits a intellectual ceiling.

​For freelancers and wordsmiths worried about their futures, this trend highlights a unique plot twist. The very technology threatening to disrupt the writing industry is currently one of the biggest employers of writers in the world.

​It turns out that in order to build the future of artificial intelligence, tech companies still have to rely on the oldest tool in human history: the human storyteller.

Does it surprise you that AI companies are still heavily dependent on human writers? Or did you see this coming? Comment below 👇