Only 18% of Americans trust AI with their data

A new survey published on April 2, 2026 by Cloaked, a consumer privacy company, finds that fewer than one in five Americans trust AI platforms to keep their personal data secure - yet nearly all of those same people continue using those platforms anyway. The findings, drawn from 1,009 U.S. adults, map a population caught between dependency and distrust, quietly deploying small acts of resistance while accepting that logging off is no longer a realistic option for most.

The report, titled "How Americans Feel About Sharing Their Data With AI," captures a population that feels simultaneously monitored and powerless. According to Cloaked, 64% of Americans believe AI is making decisions about them without their knowledge or consent. Nearly three in four respondents - 75% - say they feel constantly monitored by AI and the technology they use every day.

The trust deficit in numbers

The headline finding is stark. According to Cloaked, only 18% of Americans say they trust AI to keep their personal data secure. That means more than four out of five adults in the United States have no confidence that the AI platforms they interact with daily are handling their information safely.

That number acquires additional weight alongside a broader pattern that PPC Land has tracked across multiple research releases in 2025 and early 2026. A Shift Browser survey of 1,448 Americans published in March 2026 found that 81% were concerned about AI data access. A Usercentrics study from July 2025, covering 10,000 internet users across Europe and the United States, found that 59% were uncomfortable with their data being used to train AI systems. The Cloaked data adds a sharper point: discomfort has hardened into distrust. Concern about collection is one thing; a belief that the systems themselves are not secure is another.

The data most Americans refuse to share follows a predictable hierarchy of sensitivity. According to Cloaked, 88% of respondents said they were most uncomfortable sharing their Social Security number with AI, followed by financial information at 87% and biometric data at 74%. These three categories represent the most consequential classes of personal information - the ones most directly tied to financial identity, physical identity, and legal identity. What the survey documents is that Americans have clearly mapped out those hierarchies themselves, without prompting.

There is a structural irony here for the advertising industry. Behavioral targeting, personalization, and AI-powered ad delivery all depend on some version of the data Americans are least willing to share. PPC Land previously coveredresearch from Usercentrics showing that consumer resistance stems not from opposition to AI itself, but from insufficient transparency about how data is used. The Cloaked survey suggests that, at least for the most sensitive data categories, transparency alone may not be sufficient - the resistance is structural.

How Americans fight back

The more revealing dimension of the Cloaked research is behavioral. Americans are not passively accepting the current terms of AI data collection. They are actively, if imperfectly, fighting back.

According to Cloaked, nearly one in three Americans - 31% - have given an AI platform a fake name or fake birthday when asked for personal information. More than half, 53%, have opted out of data collection, tracking, or targeted ads at some point. And nearly two in five - nearly 40% - are more concerned about AI collecting their children's data than their own.

Physical countermeasures are also documented. Women are more likely than men to have physically covered or disabled a camera on a device due to AI privacy concerns - 43% of women compared to 34% of men. That figure is not trivial. It describes a population taking hardware-level action against digital surveillance, going beyond software settings or cookie preferences to physically modify their relationship with connected devices.

The opt-out behavior has direct relevance for programmatic advertising. More than half of Americans actively opt out of tracking where it is offered. For publishers and advertisers, the consent signal landscape is becoming more contested, and the Cloaked data reinforces what multiple pieces of research published in 2025 have documented: the audience is not passive.

The dealbreakers

Not all privacy concerns are equal. According to Cloaked, the single most cited dealbreaker among those tested - the scenario most likely to make Americans leave a platform entirely - is discovering that AI was making decisions about them regarding credit, hiring, or insurance without their consent. More than two in five Americans said that scenario would cause them to quit a platform. That exceeds even data sharing with government agencies, which 41% said would cause them to leave.

These are not abstract concerns. Automated decision-making in credit scoring, insurance pricing, and hiring is not hypothetical. It is current practice at scale across multiple industries. The survey documents that the American public is aware of this, and that awareness - rather than softening resistance - is sharpening it.

Gender and age differences are visible across these dealbreakers. Boomers are the most likely generation to say they would leave a platform if their data was shared with government agencies - 51% compared to 37% of Gen Z. But Gen Z shows its own particular form of anxiety: 59% of Gen Z respondents say they feel powerless to protect their personal data from AI, compared to 45% of Boomers. Gen Z is also the generation most likely to report that its trust in AI platforms has decreased in the past year - 28% say this, compared to just 11% of Boomers.

The economic dimension is also documented. According to Cloaked, 44% of Americans say they would pay more for a service that guaranteed their personal data would never be processed by AI. Among Gen Z, that figure rises to 49%. Among Boomers, it falls to 37%. The willingness-to-pay signal is commercially significant. It suggests a market for privacy-preserving services that is meaningful in size and skews toward a younger, digitally-active population.

Which platforms earn the most trust

The Cloaked survey ranked AI platforms by the comfort their users express with AI handling personal data, and the results produce a hierarchy that cuts against some assumptions.

According to Cloaked, Meta AI users are the most comfortable of any group with AI handling their personal data, at 67%. Claude users rank second at 63%. Grok users rank third at 62%. When the question shifts to trust specifically - which users are most likely to trust AI to keep their data secure - Meta AI users again lead at 36%. Perplexity users rank second at 31%. Grok users rank third at 27%. ChatGPT users, despite representing a large share of the overall AI user population, trust AI to keep their data secure at just 20% - roughly half the rate of Meta AI users.

These rankings sit within the dataset of the Cloaked survey and should be read as reflecting the responses of that survey's participants rather than as a market-wide measurement of platform trust. The sample methodology covers 1,009 U.S. adults, with a generational breakdown of Gen Z at 17%, Millennials at 52%, Gen X at 23%, and Baby Boomers at 7%.

That said, the result for Meta AI is counterintuitive given the backdrop. PPC Land has covered extensively how Meta announced in April 2025 that it would use public posts and comments from users over 18 to train its AI models, prompting European regulatory scrutiny and a German court ruling in August 2025 confirming that children's data was included in AI training despite announced age restrictions. A separate noyb-commissioned survey found that only 7% of people wanted Meta to use their data for AI. The Cloaked finding does not contradict those results - it reflects a different question. Among people who are already Meta AI users, comfort with the platform is higher than for other platforms. Platform loyalty, familiarity, and integration into existing social media habits may all contribute to that result.

What scenarios feel most invasive

The survey also mapped which AI-adjacent scenarios respondents find most intrusive. According to Cloaked, nearly three in four Americans - 73% - say a bank reviewing their social media posts to determine creditworthiness feels creepy or invasive, with Boomers the most alarmed at 82%. More than two in three - 70% - say an app that listens to their conversations to recommend products feels creepy.

These two scenarios are structurally different but both involve AI inference operating beyond the immediate, visible transaction. The bank scenario involves using data from one context - social media behaviour - to make decisions in an entirely different context - lending. The listening app scenario involves passive, ambient collection. Both represent forms of data use that consumers have not explicitly authorised, and both are regarded with deep discomfort by substantial majorities.

There is a partial exception. According to Cloaked, 52% of Americans feel comfortable with AI detecting fraudulent transactions. That compares with 59% who feel uncomfortable with AI scanning their emails to personalise ads. The contrast is instructive. Americans are not opposed to AI operating on their data in principle - they are opposed to it when the primary beneficiary is the platform rather than the user.

That distinction matters for digital advertising. The advertising model requires AI to operate on user data in ways that benefit the platform and its advertisers. The Cloaked survey suggests that Americans have internalised the difference between protective AI use and extractive AI use, and that the latter - which encompasses most of what underpins the programmatic ecosystem - faces significant public resistance.

Control, resignation, and the limits of opt-out

A striking feature of the Cloaked data is how it documents the gap between awareness and action. Nearly two in three Americans say they have less control over their personal data today than five years ago. More than one in two say they have accepted that companies know more about them than they are comfortable with.

Boomers are the most likely generation to believe some loss of privacy is an unavoidable cost of modern technology - 65% hold that view, compared to 54% of Gen Z. The generational contrast here is revealing: younger Americans are more likely to feel powerless, but less likely to have accepted the situation as permanent. That combination - heightened distress without resignation - may translate into more active resistance from Gen Z over time.

Gen X shows the strongest deletion behaviour. According to Cloaked, 56% of Gen X respondents have deleted an account or app due to AI privacy concerns, compared to 42% of Millennials and 43% of Gen Z. Gen X's higher deletion rate may reflect both greater seniority in digital services - more accounts to delete - and a generational cohort that came of age before surveillance capitalism was normalised.

For the advertising and marketing industry, the broader context here is the structural tension that PPC Land has documented through 2025 and into 2026: AI adoption is accelerating, the trust infrastructure supporting it is thin, and the gap between usage and confidence is not narrowing. The Cloaked survey, published on April 2, 2026 and covering data collected from 1,009 U.S. adults, adds another layer of specificity to that picture. Americans are not choosing between using AI and protecting their privacy. They are doing both simultaneously, imperfectly, and with a clear-eyed sense that neither effort is fully working.

Timeline

July 1, 2025 - Usercentrics publishes the State of Digital Trust 2025 report, finding 59% of 10,000 consumers uncomfortable with data used for AI training, covered by PPC Land
April 7, 2025 - Meta announces updated privacy policy allowing use of public posts for AI model training, with a deadline for users to object by May 27, 2025, as covered by PPC Land
August 7, 2025 - noyb-commissioned survey finds only 7% of people want Meta to use their data for AI, covered by PPC Land
August 17, 2025 - German court confirms Meta AI training includes children's data despite announced age restrictions, as reported by PPC Land
November 21, 2025 - Gmail issues public clarification that it did not change users' AI data settings, covered by PPC Land
December 7, 2025 - Verve survey documents 65% of consumers worried about AI data training, 97% demanding greater transparency, as reported by PPC Land
March 3, 2026 - Shift Browser publishes 2026 AI Consumer Insights Survey finding 81% of Americans concerned about AI data access, covered by PPC Land
March 10, 2026 - Spain's AEPD fines Yoti 950,000 euros over biometric data and consent failures, reported by PPC Land
April 2, 2026 - Cloaked publishes survey of 1,009 U.S. adults finding only 18% trust AI to keep personal data secure, with Meta AI users ranking highest on comfort and trust across AI platforms

Summary

Who: Cloaked, a consumer privacy company, surveyed 1,009 U.S. adults across generational groups including Gen Z (17%), Millennials (52%), Gen X (23%), and Baby Boomers (7%).

What: The survey finds that only 18% of Americans trust AI to keep their personal data secure. Nearly one in three admit to providing fake personal information to AI platforms. Meta AI users are the most comfortable with AI handling their personal data at 67%, followed by Claude users at 63% and Grok users at 62%. The single most cited dealbreaker for platform loyalty is discovering that AI is making consequential decisions - about credit, hiring, or insurance - without user consent.

When: The survey was published on April 2, 2026, covering responses collected from 1,009 U.S. adults.

Where: The research was conducted in the United States, covering a nationally representative adult population across generations and demographic groups.

Why: As AI platforms become more deeply integrated into daily life - and as the advertising and data industries build larger systems on top of AI-processed consumer data - the Cloaked survey documents the depth and specificity of public distrust. The findings are relevant to marketers, publishers, platform operators, and regulators who need to understand not just how many people are worried, but which data types, which scenarios, and which demographic segments carry the highest resistance.