Zappi study: Consumers open to AI ads, but trust and personalization risks remain

As OpenAI rolls advertising into ChatGPT, new research from consumer insights platform Zappi suggests consumers are not automatically resistant to advertising inside AI assistants — but trust becomes fragile when ads interfere with recommendations, personalization, or sensitive conversations.

Zappi surveyed 1,000 U.S. adults who currently use AI assistants including ChatGPT, Gemini, Claude, Copilot, and Perplexity. These are active users, not casual observers:

• 56% use AI assistants daily or multiple times per day

• 61% have used ChatGPT in the past six months, followed by Gemini (41%), Meta AI (25%), and Copilot (18%)

• 31% currently have access to premium AI subscriptions, either personally (22%) or through work or school (9%)

The report, Lines in the Sand: Where Consumers Will, and Won’t, Accept Ads Within AI Assistants, explores how early adopters think about trust, switching behavior, personalization, and acceptable contexts for advertising inside AI assistants.

Consumers are not reacting to AI advertising with a simple yes-or-no answer. Instead, they’re splitting into distinct camps.

Among free AI assistant users, 60% say they would continue using free versions supported by ads. These users would be equally likely to pay for an ad-free version (17%) as they would to switch to another AI assistant without ads (17%). To a large contingent of this group, advertising on AI assistants is acceptable as long as the tools remain free (24%). 

Among paid users, more than 45% say they would continue paying to avoid ads, while 20% say they would consider switching platforms entirely.

Consumers also appear willing to continue down the purchase funnel from AI ads. 50% say they would likely research a product further after seeing a relevant ad inside an AI assistant, while 43% say they would likely click through to an advertiser’s website.

Consumers are less worried about seeing ads than they are about what ads might do to the integrity of AI responses.

42% cite “too many ads” as a concern, but many of the strongest reactions center on trust and objectivity:

• 33% worry AI responses will become biased toward advertisers

• 32% cite privacy and data concerns

• 28% worry ads will become difficult to distinguish from real answers

• 27% worry about commercial influence over recommendations

Just 13% say they have no major concerns about advertising inside AI assistants.

At the same time, 82% say AI ads would feel at least as trustworthy as Google Search ads, including 33% who say they would feel more trustworthy.

Younger users are especially receptive. 42% of 18–34-year-olds say AI ads would feel more trustworthy than search ads — more than triple the rate among users 55–75.

But consumers draw sharp boundaries around where ads become unacceptable:

• 17% say ads would be most unacceptable while working on confidential professional documents

• 16% point to mental health conversations

• 14% point to medical questions

• 12% point to financial or legal discussions

The challenge for AI platforms may not be whether consumers accept ads at all — but whether platforms can avoid violating trust in the moments that matter most.

52% of respondents say they would change how they use memory and personalization features if AI assistants introduced advertising.

27% say they would become less willing to use personalization entirely. Another 25% say they would keep memory features enabled but opt out of ad targeting specifically.

The findings point toward a growing tension at the center of AI monetization: the advertising systems platforms want to build may discourage the very data-sharing behavior those systems depend on.

“AI advertising is arriving at the most consequential moment in the technology’s adoption curve,” said Nataly Kelly, CMO at Zappi. “The consumers in this survey are the people who set the norms that the mainstream inherits. What they find acceptable, and where they draw lines, will define the channel for years. It’s so important that the platforms – and the advertisers – understand the context and what’s required from consumers to trust the service and ads alike”

Methodology

Zappi surveyed 1,000 U.S. adults who currently use AI assistants in May 2026 using the Zappi platform. This research reflects stated attitudes among current AI assistant users and is intended to capture directional sentiment among early adopters rather than forecast precise future behavior.