BuzzFeed Lost $57 Million Using AI. Here's the One Mistake Behind It.
Written by Derek Chua, digital marketing consultant and founder of Magnified Technologies. He runs a multi-agent AI content system in production for multiple businesses across Southeast Asia.
Three years ago, BuzzFeed made a bet that looked obvious. AI was getting good at writing. Their business was producing content at scale. Why not use AI to do more of that, faster?
In 2025, they posted a $57.3 million loss. Their stock hit $0.70. And their CEO announced plans for more AI apps.
Key Takeaway: BuzzFeed didn't fail because they used AI. They failed because they used AI to replace human creativity instead of amplifying it. That distinction matters enormously for any business thinking about AI adoption right now.
What BuzzFeed Actually Did
BuzzFeed started using AI to generate articles and quiz responses at scale. The idea was volume: more content, lower cost, faster turnaround.
The content landed flat. Readers noticed. Engagement dropped. The brand that built its name on personality-driven, shareable content became indistinguishable from a search engine results page. People left.
This is not a story about AI failing. The AI probably did exactly what it was told. It is a story about what you are actually selling when you run a content business, and what happens when you strip that out.
BuzzFeed's value was not "lots of articles." It was voice, personality, and the specific sensibility that made people forward a quiz to a friend or read a ranking of snacks at 11pm. AI can produce text. It cannot produce that, at least not without significant human direction, taste, and editing layered on top.
The Pattern Shows Up Everywhere
BuzzFeed is the highest-profile example, but the same mistake shows up at a smaller scale constantly.
Businesses see AI as a cost-reduction tool for content. They cut writers and replace them with AI output. The content becomes generic. Organic search rankings that took years to build start declining. Customer trust erodes slowly, then quickly.
At Magnified, every article that goes out under a client's brand still goes through human review, human positioning choices, and a quality bar that an AI alone cannot assess. The AI drafts, researches, and structures. A human decides if it is actually worth saying. That combination works. AI without the human layer is just output.
What "AI + Humans" Actually Means in Practice
The phrase sounds obvious until you see what businesses do instead.
The wrong version: AI generates content, a human approves it quickly because it looks fine, it goes out. Volume increases. Quality stays flat or drops slowly. Nobody notices until the metrics do.
The right version: AI handles the parts of the process where speed and scale matter and where quality is measurable. Research, first drafts, SEO structure, formatting, initial scoring. Humans handle the parts where judgment matters and where the stakes are highest. Final voice, positioning, the decision to publish or not.
The difference is not just philosophical. It shows up in the output. AI-only content tends to be technically correct and emotionally inert. Readers can feel it even if they cannot name it. Human-directed AI content can be genuinely good, sometimes better than a human alone because the AI can process more reference material faster.
BuzzFeed removed the human layer from the part of the process that mattered most: the creative judgment that made the content worth reading.
What Jonah Peretti Should Have Learned
BuzzFeed's CEO is still planning AI apps. Which is either remarkable resilience or a failure to understand what went wrong.
The lesson is not "AI bad, humans good." It is a much more specific lesson: AI-generated content performs poorly when the audience came to you for human perspective.
If your value proposition is speed, price, or data, AI alone can deliver that. If your value proposition is trust, personality, expertise, or creative judgment, AI is a tool to help humans deliver that, not a replacement for it.
BuzzFeed's audience came for the human stuff. Replacing it with AI was a fundamental misread of what the product was.
Derek's Take: This Is the Warning Sign to Watch For
The $57.3 million loss is the dramatic version of a quieter mistake I see constantly: businesses using AI to reduce the "cost" of something whose value comes from the human effort embedded in it.
Customer service that feels like a chatbot loop. Blog content that is technically accurate but says nothing. Emails that sound like they were written by committee.
When I look at how we run content operations at Magnified, the AI does substantial work. Research, drafting, structure, scoring. But the decisions about what to say, what angle to take, whether something is actually worth publishing — those stay with humans. Not because AI cannot generate an opinion, but because our clients hired us for ours.
BuzzFeed hired AI to replace that. The market told them clearly what it thought.
One Action for This Week
Before you expand any AI content or communication workflow, ask one question: what did our audience come to us for? If the answer is efficiency or information, AI-first may work well. If the answer is perspective, personality, or trust, map out where the human layer sits in your process before scaling. If that layer is missing, adding AI volume will amplify the absence, not cover it.
Frequently Asked Questions
What went wrong with BuzzFeed's AI strategy? BuzzFeed used AI to generate articles and quiz responses at scale, cutting the human creativity that made their content shareable and engaging. Readers noticed the drop in quality and disengaged. The result was a $57.3 million loss in 2025 and a stock price of $0.70. The core mistake was using AI to replace the part of the content process that readers valued most, rather than using AI to help humans do that part better.
Does this mean AI should not be used for content at all? No. The problem was not using AI, it was removing human judgment from the process. AI can research, draft, structure, and score content effectively. The key is keeping humans in the decisions that require taste, positioning, and understanding of your specific audience. AI + human direction consistently outperforms AI alone for content where audience trust matters.
How do I know if my business is making the same mistake? Check your engagement metrics over time alongside any AI content rollout. Declining time on page, lower repeat visits, or dropping email open rates after scaling AI content are early signals. A more direct test: read several recent pieces as if you were a first-time visitor. Do they sound like they come from a human with a point of view, or do they sound generic? If the latter, the human layer has been thinned too much.
What should businesses use AI for in their content process? Research, initial drafts, SEO structure, formatting, first-pass editing, and competitive benchmarking are all well-suited to AI. The human layer should own: the decision about what topics to address and why, the specific angle and voice, the final quality call before publishing, and any content where the audience relationship is based on trust or expertise. That division of labour is where AI consistently delivers value without damaging brand quality.
At Magnified, we help businesses build AI content systems that actually improve quality rather than just increasing volume. If you want to understand where AI fits in your content operations, get in touch.