Your AI assistant just became an ad platform.

Welcome back.

Quick note before we dive in: if someone forwarded this to you, you can subscribe free at wiredonai.ai.

Let’s get into it.

OpenAI launched a self-serve advertising platform for ChatGPT this month. Any business can now buy ads that show up inside your conversations.

Read that again. The tool you use to think, write, research, and make decisions is now selling your attention to advertisers.

How it works

When you ask ChatGPT something, ads can now appear in labeled boxes at the bottom of its responses. OpenAI says the ads are “contextual” — matched to what you’re talking about, your chat history, and your previous ad interactions. They don’t influence the actual answer. That’s the claim.

The self-serve Ads Manager uses CPC and CPM bidding — same model as Google Ads. No minimum spend, so every business from Fortune 500s down to local shops can buy in. Retailers can even connect product feeds so ChatGPT automatically generates ads from their catalogs.

OpenAI’s revenue goal: $2.5 billion in ad revenue this year. $100 billion by 2030.

Why this should make you uncomfortable

I’m not going to pretend this is fine just because the ads are “clearly labeled.”

The whole reason people switched to ChatGPT from Google was to get away from the ad-cluttered search experience. You ask a question, you get an answer — no sponsored results, no SEO games, no wondering if the top recommendation is there because it’s best or because someone paid for the placement.

That era is now officially over.

OpenAI says ads won’t influence the AI’s actual responses. But here’s the problem: even if that’s technically true today, the incentive structure has permanently changed. When your revenue depends on advertisers being happy, every product decision gets filtered through “will this hurt ad revenue?” That’s not a conspiracy — it’s how every ad-supported platform has worked since the invention of television.

Google’s search results didn’t get worse overnight. It happened gradually, over years, as the ad business grew and the line between organic and sponsored got blurrier. There’s no reason to think ChatGPT will be different.

What this means for you

Three things to keep in mind:

First, start treating ChatGPT’s product recommendations the way you treat Google’s. If you ask “what’s the best project management tool?” and it suggests something, you now have to wonder whether that answer was influenced by advertising.

Second, if you’re paying $20/month for ChatGPT Plus, you should be asking why you’re also seeing ads. Historically, subscription products have been ad-free. OpenAI hasn’t confirmed whether Plus subscribers will see ads or not, but the fact that they haven’t ruled it out tells you something.

Third, this is the clearest argument yet for having alternatives. Claude doesn’t run ads. Gemini doesn’t run ads (yet). Local models running on your own hardware never will. The more you diversify which AI you use for what, the less exposed you are to any single platform’s incentives changing under you.

My take

I’ve been saying for a while that the AI tools we rely on will eventually have to pick a side: are you serving the user, or are you serving the advertiser? OpenAI just picked. They want to be both. History says that doesn’t work out well for the user.

The practical move: pay attention to which AI tools are ad-supported and which aren’t. Make that part of your evaluation. Right now, the best AI writing and analysis tool (Claude) happens to also be the one with no ads. That might not last forever, but it’s worth noting.

Claude just passed ChatGPT in business adoption. For the first time ever, more American businesses are paying for Anthropic’s Claude than OpenAI’s ChatGPT — 34.4% vs 32.3%, according to Ramp’s spending data. Over the past year, Anthropic quadrupled its business adoption while OpenAI grew just 0.3%. The biggest driver? Claude Code, Anthropic’s AI coding tool, which now authors an estimated 4% of all public GitHub commits worldwide. Businesses aren’t switching because of hype — they’re switching because of output quality.

Anthropic is now worth almost a trillion dollars. The company is in talks to raise $30-50 billion at a $950 billion valuation — higher than OpenAI’s last valuation of $852 billion. That’s a company that didn’t exist four years ago. Whether that valuation makes sense depends entirely on whether they can keep shipping better products than the competition.

Coinbase cut 14% of its workforce because of AI. About 700 people. CEO Brian Armstrong said engineers are now shipping in days what used to take teams weeks. They’re replacing “pure managers” with “player-coaches” who both manage and build. They’re also creating “AI-native pods” — one-person teams directing AI agents that handle what used to require separate engineers, designers, and PMs. This is the playbook more companies will follow. Whether your job is safe depends on whether you’re the person using AI to multiply your output, or the person whose output AI just replaced.

Google announced the “Googlebook.” A new line of laptops that merges Android and ChromeOS into one operating system. More details expected at Google I/O on May 19. If this works, it’s Google’s answer to Apple’s tight hardware-software integration — but with AI baked deeper into the OS layer.

Now that AI tools are starting to carry ads, here’s a prompt to pressure-test any recommendation you get:

I’m going to ask you to recommend [a tool / product / service] for [my specific use case]. Before you answer, I want you to:

Now: what’s the best [tool/product/service] for [your specific need]?

Most AI tools will say they don’t have commercial relationships. That’s fine — the point is training yourself to ask the question. The day the answer changes, you want to already be in the habit of checking.

Enjoying Wired on AI? Forward it to someone who’d find it useful. It’s free to subscribe at wiredonai.ai.

Until next week,

— Scott

P.S. — Are you bothered by ads in ChatGPT, or is it just the cost of free AI? Reply and tell me — I want to know where people draw the line.