The Original Sin of the Internet Is Now Inside Your AI

The News

OpenAI Just Opened the Floodgates

We've seen this before. A transformative technology emerges, promising to democratize information and empower people. Early adopters love it. Then the incentive structure shifts. Then the ads come. Then the product slowly, quietly, reliably degrades into something that serves shareholders first, advertisers second, and users last.

We watched it happen to search. We watched it happen to social media. We watched it happen to news, to email, to apps. Now we're watching it happen — in real time, in slow motion — to AI.

On May 5, 2026, OpenAI opened its self-serve advertising platform to the public. Any business can now buy ads inside ChatGPT conversations on a cost-per-click basis, with no minimum spend required. Best Buy, Lowe's, Target, Albertsons, and Williams-Sonoma were among the first pilot brands. Agency partners include Dentsu, Omnicom, Publicis, and WPP. The ad tech ecosystem — Adobe, Criteo, Kargo, StackAdapt — is already plugged in.

The pilot hit $100 million in annualized recurring revenue in under two months. OpenAI is just getting started.

The Pattern

We've Seen This Movie Before

In 1998, Google launched with a simple promise: return the most relevant result. There were no ads. In fact, the founders — Sergey Brin and Larry Page — actually wrote a research paper warning that ad-supported search engines were inherently biased. The financial incentives, they argued, would inevitably corrupt the results.

Within four years, Google had launched AdWords. By 2010, advertising was the company's core business. By 2024, Google generated $264 billion in advertising revenue — roughly 77% of Alphabet's total revenue. Today, type any commercial query into Google and count the ads before you reach an organic result. There are routinely four or more above the fold.

Meta's arc is even starker. Mark Zuckerberg famously said Facebook was about "connecting the world," not making money. Now 97% of Meta's revenue — $160 billion in 2024 — comes from advertising. The product still claims to connect people. What it actually does is auction their attention to the highest bidder, 24 hours a day.

Cory Doctorow named this pattern "enshittification" — the inevitable three-stage decay of digital platforms: first they're good for users, then they abuse users to serve advertisers, then they abuse both to maximize shareholder returns. The American Dialect Society named it Word of the Year for 2023. By then, everyone knew exactly what it meant.

This is the pattern OpenAI is now entering. Not by accident. Not despite the founding principles. But as the explicit, investor-approved, boardroom-blessed strategy for reaching a $100 billion ad empire by 2030.

The Danger

But This Time, the AI Knows You

Here's what's different about advertising inside an AI assistant — and it matters even if you've made your peace with ad-supported search.

Google knows what you searched for. ChatGPT knows what you're thinking about.

When you type "best running shoes" into Google, you reveal a purchase intent. When you ask ChatGPT "I've been having knee pain after my runs — what might be causing it, and should I keep training for my marathon?" you reveal something categorically more intimate. A fear. A goal. A physical vulnerability. A timeline. A commitment. You're not querying a database. You're thinking out loud to something that responds, remembers, and adapts.

ChatGPT now has 900 million weekly active users — more than double what it had in February 2025. OpenAI projects 2.75 billion weekly users by 2030. That's roughly the entire current internet, pouring their most private questions, anxieties, and plans into a single product owned by a company that just decided advertising is its future.

The Fork

The Fork in the Road

This same week, Anthropic — OpenAI's closest competitor and a company founded largely by former OpenAI researchers — made a different announcement. Claude will never show ads. Full stop.

Anthropic published a statement explaining that an advertising-based business model introduces incentives that work against genuine helpfulness. "Unlike a list of search results," the company wrote, "ads that influence a model's responses may make it difficult to tell whether a given recommendation comes with a commercial motive or not. Users shouldn't have to second-guess whether an AI is genuinely helping them or subtly steering the conversation towards something monetizable."

The timing — Anthropic even ran a Super Bowl ad in February staking its identity on being ad-free while OpenAI's pilot was already running — is almost too on the nose. Two companies that share a founding lineage, two visions of what AI is actually for.

This is a genuine fork — not just for two companies, but for the AI industry's relationship with everyone who uses it. Which model wins depends on whether people actually care, or whether they're willing to accept sponsored answers in exchange for a free chatbot, the same way they accepted sponsored results in exchange for free search.

The history of the internet suggests they will. Most people didn't leave Google when the ads came. Most people didn't leave Facebook when the feed became an auction. Free is a powerful force. And OpenAI is offering a free tier. The ads are on the free tier. The math is familiar.

The Irony

The Founding Irony

OpenAI was incorporated in 2015 as a nonprofit explicitly designed to avoid the distorting effects of profit-seeking on AI development. The founding logic was clear: if you build AGI under shareholder pressure, you make shortcuts. You move too fast. You deploy things before they're ready. You build things that serve revenue rather than humanity.

Now, ten years later, OpenAI is the world's most powerful AI company — backed by Microsoft, valued at over $300 billion — and it's building what could become the largest advertising platform in history. Inside the most intimate digital interface most people have ever used. Sold to major agencies. Running on cost-per-click bidding. Targeting 2.75 billion users by the end of the decade.

The tension between "we're building AGI safely" and "here's a sponsored answer" isn't a contradiction that will resolve itself cleanly. It's a fault line that will quietly shape every product decision, every model deployment, every conversation about what ChatGPT is actually for.

We've watched ads hollow out search. We've watched them hollow out social media, news, podcasts, and email. At each step, the product got a little worse — a little more optimized for engagement over accuracy, a little more interested in keeping you clicking than in giving you what you actually needed.

The AI revolution was supposed to be different. For a moment, it was. The question now is whether that moment is already over — or whether the fork Anthropic just identified is real enough, and whether enough people will choose it to matter.

---

Watch what happens next. Does the quality of ChatGPT's free-tier responses start to drift in subtle, hard-to-prove ways? Do users notice? Does the press hold OpenAI accountable to its original mission? Does Anthropic's bet on trust become a competitive advantage — or does "free" win again, the way it always has?

The original sin of the internet replicated itself inside every major digital platform of the last 30 years. The bet OpenAI is making is that history will repeat — and that this time, the platform it happens on knows you more intimately than any that came before it.