Why DeepMind’s CEO Is Surprised by OpenAI’s ChatGPT Ad Push
Google DeepMind CEO Demis Hassabis has voiced surprise over OpenAI’s decision to begin testing ads inside ChatGPT—a move that could reshape how users interact with AI assistants. Speaking at the World Economic Forum in Davos, Hassabis questioned the timing and philosophy behind inserting advertising into a tool designed to act as a personal, trustworthy aide. With ChatGPT boasting 800 million weekly active users, many of whom rely on its free tier, the ad rollout marks a pivotal moment in the commercialization of generative AI.
Hassabis didn’t condemn ads outright—in fact, he acknowledged their role in funding much of today’s internet—but he raised deeper concerns about user trust, assistant neutrality, and whether monetization should come before maturity in AI systems.
The Timing of Ads Raises Eyebrows in the AI Community
OpenAI’s announcement that it would test ads within ChatGPT came amid mounting pressure to offset ballooning infrastructure and energy costs. As large language models grow more powerful, they also become exponentially more expensive to run. For a company still navigating profitability despite massive user engagement, ads may seem like a logical next step.
Yet Hassabis sees it differently. “I’m a little bit surprised they’ve moved so early into that,” he told Axios during a Davos interview. His comment reflects a broader tension in the AI industry: how fast should companies monetize tools that are still evolving—and still prone to errors?
Unlike traditional software products, AI chatbots are increasingly positioned as personal agents—tools that users confide in, delegate tasks to, and even form emotional attachments with. Introducing ads at this stage, Hassabis suggests, risks undermining the very relationship these platforms are trying to build.
Trust vs. Revenue: The Core Dilemma for AI Assistants
At the heart of Hassabis’s concern is a philosophical question: Can an AI assistant truly serve your best interests if it’s also trying to sell you something?
“If you think of the chatbot as an assistant that’s meant to be helpful—and ideally, in my mind, as they become more powerful, the kind of technology that works for you as the individual…there is a question about how ads fit into that model?” he asked.
This isn’t just about banner placements or sponsored links. In an AI context, ads could subtly influence recommendations, search results, or even conversational tone. Imagine asking for restaurant suggestions and receiving only those from paid partners—or seeking health advice and being nudged toward promoted supplements. Even if disclosed, such dynamics erode neutrality.
For Google, which built its empire on search ads, the temptation must be strong. Yet Hassabis emphasized that DeepMind and Google have “no current plans” to introduce ads into their AI chatbot. Instead, they’re taking a “very careful” approach, prioritizing user experience over short-term revenue.
Why Google Is Holding Back—For Now
Despite Google’s deep reliance on advertising revenue across its ecosystem, the company appears unwilling to rush ads into its AI assistant. Hassabis stressed that there’s no internal pressure to make “a knee-jerk” decision, even as competitors move faster.
Injecting ads too soon into an AI assistant could compromise perceived trustworthiness, especially if users feel manipulated or misled.
Moreover, Google’s AI offerings, including Gemini, are still being refined for accuracy, safety, and reliability. Leadership seems to believe that establishing a solid foundation of user trust must precede any aggressive monetization strategy.
“We’re monitoring the situation,” Hassabis said, signaling that while ads aren’t off the table forever, they won’t be introduced until the team is confident they won’t harm the assistant-user relationship.
What OpenAI’s Move Means for the Future of AI Monetization
OpenAI’s ad experiment could become a defining case study in AI business models. On one hand, ads offer a scalable path to sustainability—especially as compute costs soar and investor expectations intensify. On the other, they risk alienating the very users who fuel network effects and data feedback loops.
Notably, OpenAI isn’t abandoning subscriptions; its premium tiers (like ChatGPT Plus) remain central to its revenue mix. But with only a fraction of users paying, ads may be seen as necessary to support the free tier that drives mass adoption.
Still, the optics matter. Rolling out ads before fully solving issues like hallucination, bias, or inconsistent reasoning could backfire. Users might start questioning whether responses are shaped by algorithms—or advertisers.
Industry observers note that if OpenAI can implement ads transparently and non-intrusively—perhaps through clearly labeled sponsored answers or opt-in promotions—it might set a new standard. But if the execution feels exploitative, it could open the door for rivals like Google to position themselves as the “trust-first” alternative.
User Experience Hangs in the Balance
Mobile users, who make up a significant portion of ChatGPT’s audience, are especially sensitive to cluttered interfaces and unexpected interruptions. In 2026, with mobile readability and seamless UX being critical ranking factors for Google Discover and search visibility, any ad integration must be frictionless.
Hassabis’s hesitation likely stems from real-world lessons: past tech giants have lost user goodwill by prioritizing monetization over utility. Remember when social feeds became unrecognizable due to algorithmic ads? Or when “helpful” pop-ups turned into conversion traps?
AI assistants operate in a more intimate space—they’re not just apps; they’re conversational partners. That intimacy demands higher standards. As Hassabis put it, “You want to have trust in your assistant.” Once that trust is broken, it’s hard to rebuild.
A Strategic Crossroads for the AI Industry
The divergence between OpenAI and Google highlights a strategic fork in the road for AI development. One path prioritizes speed-to-revenue, betting that users will tolerate ads if the core experience remains valuable. The other prioritizes long-term trust, banking on delayed monetization to build a more loyal, engaged user base.
Neither approach is inherently wrong—but the stakes are high. Whichever model proves more sustainable could shape the next decade of human-AI interaction.
For now, Hassabis and DeepMind are choosing patience. “There’s nothing wrong with ads,” he reiterated, “if done well, they can be useful.” But in the realm of personal AI, “done well” means more than just technical execution—it means respecting the user’s autonomy, attention, and intelligence.
As the ad tests roll out in ChatGPT over the coming weeks, all eyes will be on user reaction. Will people shrug it off as the cost of free access? Or will they start looking for alternatives that feel less commercialized?
One thing is clear: in the race to dominate AI, trust may be the most valuable currency of all.
