Adblock Plus has released a new report examining the growing presence of advertising within AI services such as ChatGPT and Microsoft Copilot. The research found that 72% of U.S. adults are concerned about private AI conversations being used for ad targeting. Experts cited in the report also raised concerns that advertising within AI experiences may be difficult for consumers to distinguish from other content, potentially creating confusion around recommendations and responses.
Among the report’s key findings:
- Nearly half (49%) of internet users have used AI to find information in the past month, equivalent to around 1.59 billion people, and almost a quarter (23%) say AI is their main source when researching brands, products, or services.
- AI companies turn to advertising as one of the business models to fund these services, as an estimated $7.6 trillion is needed to scale the compute, data centers, and power required to fuel AI, making advertising an increasingly important business model.
- 72% of US adults are concerned about AI assistants using their conversation data to target them with ads, while 44% report being “very” or “extremely concerned.”
- When asked about their reaction to ads inside AI assistants, users that are concerned mention they find those annoying, intrusive and that ads make them trust the answers of an AI less.
- Adblock Plus and AdBlock have blocked more than 25 million ads on ChatGPT from April 30 tilll now and over 2,7 million ads in the last week alone.
Researchers warn that the hardest ads to spot may still be coming: academics who study advertising in conversational AI tested ads within the answer text itself. Such ads change how ad delivery works, can be generated on the fly, and are far harder for users to recognize.
“AI assistants and chatbots are becoming central to people’s lives and how they get information, but we know almost nothing about the ads inside these tools: how they work, how many there are, how they are chosen and, most importantly: what users think about ads inside AI”, said Cornelius Witt, director of global public affairs at Adblock Plus. “We want to open this black box and our report is a first look inside the new AI ad economy and what it means for users”.
What the ads look like
Through tests of ChatGPT and Copilot in August 2026, the report documents very different approaches. ChatGPT places an advertiser’s own creative in a labeled box after the answer. Copilot generates its own ad copy, in the assistant’s voice, tailored to the user’s prompt, and places it above the “Sponsored” label disclosing ads.
In both cases, the report found that the ad was matched to the prompt and overall context and not related to the answer or recommendation itself. Sometimes the advertised brand was the one the assistant had just recommended, sometimes it was a competitor.
Where this is heading
“One of the most interesting takeaways from our research is that injecting advertising into chatbot responses fundamentally changes how ad delivery works. On the web, advertisers prepare assets in advance, and targeting selects which one a user sees. In AI chatbots, targeting still works in similar ways, but those assets are no longer needed. The system only needs a keyword and product matching prompt”, said Dr. Florian Schaub, University of Michigan School of Information. “I’m really curious to see how far these systems will go in using their own inference capabilities to target ads next.”
The full report is available at https://adblockplus.org/adblock-plus-ads-in-ai-report-2026.pdf.
Visit https://adblockplus.org/ for more information.
Related News:
Netwrix Expands Microsoft Cloud Security for AI Agent Identities
Commvault Cloud Rewind: Expands AI and Cloud Application Resilience
