Skip to main content

BuzzFeed’s first AI-generated articles are ad-lib quizzes

BuzzFeed’s first AI-generated articles are ad-lib quizzes


The website announced last month that it would use OpenAI tools to assist with content creation.

Share this story

An illustrated robot winking logo
Image: BuzzFeed

BuzzFeed’s first quizzes that integrate AI writing tools are live today, with Valentine’s Day-themed content like, “Date Your Celeb Crush With The Magic Of AI” and “This AI Quiz Will Write A Rom-Com About You In, Like, Less Than 30 Seconds.” The result is a slightly more interesting Mad Libs — and a much more tedious way to use ChatGPT.

In a memo to staff last month, BuzzFeed CEO Jonah Peretti told staff the company would “lead the future of AI-powered content and maximize the creativity of our writers, producers, and creators and our business.”

The first set of quizzes prompts users to input information like names, favorite foods, or a location, and the tool spits out a personalized block of text generated using artificial intelligence. BuzzFeed says human staff write the quizzes and train the prompts and that it’s “a collaborative effort” between the staffer, the AI system, and the user. BuzzFeed appears to have put some limitations around what users can input and what the tool returns — some offensive terms have been banned, and the generated text will sometimes replace inputs with something random.

The Valentine’s Day approach is far from novel. Just yesterday, The New York Times ran something similar, feeding ChatGPT several prompts to write a valentine message.

Most of the quizzes have two bylines: “Buzzy the Robot” and a human BuzzFeed staffer. The outlet has also already started to integrate native advertising and AI-generated content: one quiz, titled “We’ll Design Your Ideal Soulmate With The Magic Of AI,” is sponsored by Miracle-Gro.

BuzzFeed’s announcement that it would begin using automated tools came on the heels of other digital publishers coming under scrutiny for how they deployed — and disclosed — the use of similar systems. Tech news site CNET quietly used an AI generation tool for months on dozens of stories, most of which were later found to have errors. The first article published at Men’s Health earlier this month using an AI tool also had errors that were only corrected once Futurism asked about them.

BuzzFeed spokesperson Matt Mittenthal told The Verge last month that the company didn’t have plans to use automated technology in its newsroom. BuzzFeed’s stock has nosedived since going public in 2021, and the company has done several rounds of layoffs, the most recent of which cut 12 percent of employees in December. BuzzFeed stock jumped in January after the news broke that the company would begin using automated tools.