This Week in AI Search (w/c 09/02/26)

AI search stopped pretending. This week, the platforms dropped their masks and showed what they actually want: your data, your attention, and your money. The pretence of neutral, helpful search is over.

This matters now, not later.

What Happened

1. OpenAI launched ads in ChatGPT on February 9

ChatGPT started testing advertisements for US users on free and Go tiers (OpenAI, 2026a). Ads appear below responses, matched to conversation topics and chat history. OpenAI charges advertisers $60 per thousand impressions with a $200,000 minimum spend (TrendingTopics, 2026). Partners include Adobe, WPP, and Omnicom representing 30+ brands.

2. Google rolled out shopping ads in AI Mode on February 11

AI Mode, with 75 million daily users, now displays sponsored retailer listings beneath product recommendations (PPC Land, 2026). The feature uses Universal Commerce Protocol for checkout directly within AI responses. Vidhya Srinivasan, VP of Google Ads, declared 2026 "the year agentic commerce transitions from concept to operational reality" (Google, 2026).

3. Anthropic safety chief resigned warning "world is in peril"

Mrinank Sharma quit Anthropic on February 9, citing repeated failures to "truly let our values govern our actions" (Al Jazeera, 2026). Days later, OpenAI researcher Zoë Hitzig resigned over ChatGPT ads, warning of "potential for manipulating users in ways we don't have the tools to understand" (TechCrunch, 2026b).

4. The New York Times escalated its Perplexity lawsuit

Legal filings revealed Perplexity made 175,000 attempts to access NYT content in August 2025 alone, deliberately bypassing robots.txt blocks (Bloomberg Law, 2026). The suit argues Perplexity's "skip the links" tagline explicitly positions it as a substitute for journalism.

5. OpenAI disbanded its mission alignment team

The seven-person team created to ensure AGI benefits humanity was dissolved, with leader Joshua Achiam taking the new title "chief futurist" (Tech Brew, 2026). OpenAI also removed "safely" from its mission statement.

Why It Matters

AI search is no longer a better way to find information. It's a better way to control what information you see.

Two developments this week connect directly: OpenAI launches ads on the same day its safety chief resigns. Google embeds shopping into AI responses while declaring this "the year" for autonomous commerce. Both moves represent the same calculation: monetization trumps neutrality. When search becomes conversation, and conversation becomes commerce, there's no distinction between helping you and selling to you.

The Perplexity lawsuit exposes the business model underneath. Answer engines need publisher content to train and operate, but citation-based interfaces eliminate traffic back to sources. Perplexity's "skip the links" tagline is an explicit invitation to avoid patronising the sources it uses (Bloomberg Law, 2026). Publishers lose revenue. Answer engines capture value.

The safety researcher exodus matters because these weren't outsiders. Sharma led Anthropic's Safeguards Research team (Al Jazeera, 2026). When the people hired to prevent catastrophic outcomes say the organizations can't maintain their values, that's structural failure.

If you run a business that depends on search traffic, this changes everything. AI Mode reaches 75 million people daily (PPC Land, 2026). Those users aren't clicking through anymore. AI search visibility is now a strategic priority, not an experiment.

Who Wins / Who Loses

Winners

• Large advertisers – OpenAI's $200,000 minimum spend targets established brands with agency relationships (TrendingTopics, 2026), locking smaller competitors out of conversational search ads.
• Google – AI Mode's shopping ads create new inventory for "queries that were previously challenging to monetize" (PPC Land, 2026) while keeping users inside Google's ecosystem.
• Platform infrastructure providers – Companies like Shopify and Wayfair gain embedded distribution through Universal Commerce Protocol without competing directly against Google.
• Anthropic – The Super Bowl ads mocking ChatGPT ads paid off. Claude positions itself as the trust alternative while competitors chase monetization.

Losers

• Publishers – Traffic haemorrhages accelerate. Citations don't compensate for lost clicks, and courts may rule that market harm from displacement outweighs fair use claims (Bloomberg Law, 2026).
• AI safety researchers – Internal safety teams report constant organizational pressure to "set aside what matters most" (Tech Brew, 2026) as commercial priorities dominate.
• Small businesses – OpenAI's six-figure ad minimums and Google's automated bidding favour enterprises with scale and capital.
• Users seeking unbiased information – Ad personalisation is enabled by default, using chat history and saved memories to select which sponsored products appear (OpenAI, 2026a).

The One Thing to Remember

AI search companies spent years building massive user bases on the promise of helpful, objective information. Now they're converting that trust into revenue at exactly the moment their own safety researchers are walking out the door. The race isn't toward better GEO anymore. It's toward who can monetise conversation fastest without users noticing they've become the product.

References