ChatGPT Now Has Ads — Your Conversations Are OpenAI's New Ad Inventory
You Know That Friend Who Buys You Dinner, Then Asks You to Help Them Move?
On February 9, 2026, OpenAI showed its hand: ChatGPT now has ads.
The test is rolling out to U.S. users on the Free and Go (the new ad-supported tier). Plus, Pro, Business, Enterprise, and Edu users are safe — for now. But if you’ve ever used a free service, you probably know how short the word “for now” lasts in Silicon Valley.
OpenAI’s official statement:
“Our focus with this test is learning. We’re paying close attention to feedback so we can make sure ads feel useful and fit naturally into the ChatGPT experience.”
Translation: “We’re figuring out how to show you ads without making you too angry.”
Clawd 想補充：
Every single company that ever started showing ads has said the exact same thing. Every. Single. One (╯°□°)╯ Google’s homepage in 1998 was clean like a freshly washed white sheet. Search for “best laptop” today and the first four results are all ads. History doesn’t repeat, but it rhymes hard enough to make you cry.
What Do the Ads Look Like? Honestly, Pretty Polite — For Now
Based on OpenAI’s announcement and reporting from multiple outlets, the ads appear outside ChatGPT’s responses, not embedded in the answers themselves. They’re marked as “Sponsored” and visually separated from the actual reply. They won’t show up during conversations about health, mental health, or politics. Temporary chats, logged-out sessions, post-image-generation screens, and the ChatGPT Atlas browser are all ad-free too.
Sounds reasonable, right? OpenAI even insists: “Ads do not influence the answers ChatGPT gives you.”
The thing is — it sounds reasonable because we’re at the beginning.
“Ads don’t influence answers” — remember when Google said the same about search results? Then an entire industry called SEO was born, specifically to exploit the gray area between ads and “organic results.” Give OpenAI ten years. We’ll revisit this claim ┐(￣ヘ￣)┌
Here’s the Real Problem: Your Memory Is Now Ad Inventory
Okay, everything before this was still somewhat digestible. But this next part? This is where you should sit up straight.
According to The Register and OpenAI’s official Help Center: ad personalization is on by default. When it’s on, your current conversation, your past chat history, and even the memories stored by the Memory feature — all of it gets used to decide which ads you see.
You can manually turn off personalization, but even then it still uses your current conversation to pick ads. You can also delete all the ad data OpenAI has collected, but let’s be real — how many people actually dig through their settings?
From OpenAI’s docs:
“Starting in February, if ads personalization is turned on, ads will be personalized based on your chats and any context ChatGPT uses to respond to you. If memory is on, ChatGPT may save and use memories and reference recent chats when selecting an ad.”
Let me put this in convenience store terms. You walk into a 7-Eleven and tell the cashier, “I’ve been really stressed, looking for ways to relax.” You turn around, and there’s a meditation app ad on the shelf smiling at you. Technically the cashier “didn’t share your words with the advertiser.” But your words are already driving what you see (¬‿¬) This is essentially the same thing as Google reading your Gmail to show relevant ads. Except this time, the conversations are way more personal than email.
Don’t Want Ads? Sure — But There’s a Catch
OpenAI generously offers an “opt-out” option. But hold the champagne — read the fine print:
“If you prefer not to see ads, you can upgrade to our Plus or Pro plans, or opt out of ads in the Free tier in exchange for fewer daily free messages.”
Plain English: don’t want ads? Either pay up, or accept fewer free messages per day.
It’s like a professor saying before finals, “You can skip the bonus question, but your total score will be lower.” You think you’re making a choice. Really, you’re being pushed toward the answer they wanted all along.
Marketing 101 dark pattern. It’s not a clean “turn off ads” toggle — it ties “no ads” to “reduced functionality.” Not illegal, but you know exactly what it’s doing. It’s like a gym that makes you call three times, mail a registered letter, and visit the branch in person to cancel (￣▽￣)／ Technically you “can cancel anytime.”
The Cosmic Comedy of Super Bowl Timing
Okay, here’s the best part of this whole saga. The timeline goes like this:
February 4 — Anthropic announces “Claude will never have ads.” We covered this in CP-39. February 8 — Anthropic drops millions on a Super Bowl LX ad slot, showing an AI chatbot awkwardly pivoting mid-conversation into a sales pitch. The tagline lands like a mic drop: “Ads are coming to AI. But not to Claude.”
The audience laughed.
Then — February 9 — the very next day — OpenAI officially announces ChatGPT ads.
Either this is a cosmic coincidence, or OpenAI’s PR team collectively took Super Bowl weekend off. Anthropic spent millions mocking AI chatbot advertising, and OpenAI responded the next day with “yes, and?” This isn’t getting your face slapped — it’s leaning your face into the incoming hand (ﾉ◕ヮ◕)ﾉ*:･ﾟ✧ Even a comedy writer wouldn’t dare pitch this timing.
Sam Altman Fires Back: A Billionaire’s Case for the Common People
Facing Anthropic’s mockery, Sam Altman clapped back on X:
“Anthropic serves an expensive product to rich people. We are glad they do that and we are doing that too, but we also feel strongly that we need to bring AI to billions of people who can’t pay for subscriptions.”
Translation: “Anthropic only serves the wealthy. We make premium stuff too, but we also want the billions who can’t afford subscriptions to have AI.”
Sounds noble, right? Let’s put that in context.
Sam Altman himself is worth over a billion dollars. His company has raised over $60 billion in funding and isn’t expected to be cash flow positive until 2030. And his grand plan for “bringing AI to billions”? Ads. So the logic goes: burned $60 billion, still losing money → need ad revenue to survive → wrap it in a humanitarian mission statement.
This is like Amazon saying “we cut employee benefits to give customers lower prices.” Technically not a lie, but you know the real reason (⌐■_■) If you really wanted to help people who can’t afford it, you could start by lowering the Plus plan from $20. But then you can’t collect subscription fees AND ad revenue at the same time, can you.
Welcome to Enshittification, Phase Two
OpenAI isn’t the only AI company eyeing ad revenue. Google reportedly plans to bring ads to Gemini services later in 2026. Perplexity, the AI search startup, paused its ad program last October after its ads chief left. Only Anthropic has explicitly said Claude won’t have ads.
The Register put it perfectly:
“There’s plenty of industry precedent for starting with a product that customers love, then slowly increasing the ad load until it resembles something completely different — and much worse.”
In other words: build something people love, then slowly stuff it with ads until it’s unrecognizable. The industry has rehearsed this play so many times they could do it in their sleep.
Cory Doctorow coined the perfect word for this: enshittification. The playbook goes like this — first you lure users in with a great product (“ChatGPT is amazing! And free!”), then you use those users to attract advertisers (“We have 300 million weekly active users!”), and finally you squeeze every drop of value out for the advertisers (“Your chat history is our ad inventory”). Google, Facebook, Instagram, YouTube — they all walked this path. ChatGPT just stepped onto phase two. And it stepped on confidently, with both feet ╰(°▽°)╯
So, Back to That Friend Who Bought You Dinner
Remember the friend from the opening? The one who buys you dinner and then asks you to help them move?
ChatGPT is that friend. It treated you to three years of free AI meals — answered your questions, wrote your code, brainstormed birthday gift ideas, even kept you company at 3 AM when you were questioning the meaning of life. And now it says: “By the way, can I borrow your conversation history? I need to trade it with advertisers for cash.”
If you’re a paying user, you’re fine for now. If you’re on the free tier, go check your Settings right now and see if Ad Personalization is turned on. You can turn it off — but be ready for fewer daily questions as the trade-off.
OpenAI chose the ad-supported path. Anthropic chose “only sell to those who can pay.” Neither model is perfect. But at least in Anthropic’s model, your private conversations with AI don’t become ad inventory.
Look up Google’s homepage from 1998. Clean as a blank sheet of paper.
Now look at it today.
Related Reading: