Usercentrics, the Munich-based consent management company, published "The Resigned Consent Report" on September 29, 2026, a 12-page spin-off of its State of Digital Trust 2026 survey of 11,000 consumers in seven markets. According to the report, only 7% of consumers are fully comfortable granting an AI system access to their data without conditions, while 16% grant that access despite being uncomfortable - a behaviour the company calls "resigned consent." Germany posts the highest rate of any market at 24%, more than double the 11% in the United Kingdom and 10% in the United States.

In Short

A company that sells cookie banners and consent tools asked 11,000 people how they feel about letting AI see their emails, files, health records and bank accounts, and found that many say yes even though they are uneasy. That matters to anyone who reads a high opt-in rate as proof that customers trust them, because the report argues a reluctant yes and a genuine one look the same on a dashboard. The numbers are the company's own, and the report leaves out several details, including the exact questions asked and how many people were surveyed in each country.

What the report measures

The document's central claim rests on a distinction that consent dashboards do not make. A person who agrees to share data can do so because the trade feels worthwhile, or because refusing feels pointless. Usercentrics labels the second group as resigned. "The person said yes because they'd already decided no wasn't a real option," the report states.

To separate the two, the survey asked respondents how they would respond to an AI assistant requesting access to six categories of data: financial accounts, files and documents, company data, health data, email and calendar, and work tools. The answers fall into several buckets, which the report labels "fully comfortable, no conditions," "some discomfort (total)," "uncomfortable, would allow anyway," and "would stop using the AI entirely." A further category, flat refusal, appears in the country and generation charts but not in the category table.

Averaged across the six categories, 58% of consumers report some level of discomfort, according to Usercentrics. That figure ranges from 53% for work tools to 64% for financial accounts. Only 7% are fully comfortable with no conditions attached. And 16% say they would grant access anyway, despite their discomfort.

The headline ratio - reluctant permission outnumbering willing permission two to one - comes from those last two figures. Sixteen divided by seven is roughly 2.3. The report claims the pattern "outnumbers willing access two to one across every category and every generation surveyed."

The category breakdown

The full table, as published in chapter one of the report:

Category Fully comfortable, no conditions Some discomfort (total) Uncomfortable, would allow anyway Would stop using the AI entirely
Financial accounts 6% 64% 14% 24%
Files/documents 6% 59% 15% 19%
Company data 8% 57% 16% 16%
Health data 8% 57% 17% 15%
Email/calendar 8% 55% 17% 15%
Work tools 8% 53% 17% 13%

The pattern runs in an unexpected direction. Financial accounts produce the most discomfort but the lowest rate of resigned consent, at 14%. Nearly a quarter of respondents - 24% - would stop using an AI product entirely rather than hand over access to their financial accounts. "When the data is sensitive, people don't grudgingly say yes. They leave," the report states.

Work tools sit at the opposite end. Discomfort there drops to 53% and abandonment to 13%, yet resigned consent holds at 17%, level with email and health data. According to Usercentrics, work tools are "where AI adoption is actually happening right now, the copilots, the assistants, the agents companies are rolling out to their own people." The company's reading: "people aren't refusing AI at work, they're just not meaning the yes."

Resigned consent peaks at 17% in three categories - work tools, email and calendar, and health data - with company data one point behind at 16%. Those four are also the categories with the lowest total discomfort. The report links this to consent fatigue, which it describes as the pattern by which "the more frequently someone gets asked the same kind of question, the less they engage actively with it." The report attributes the concept to "design researchers" without naming any study.

One structural detail is not explained in the document. In every row, the "fully comfortable" and "some discomfort" columns add up to between 61% and 70%, which leaves roughly 30% to 39% of respondents unaccounted for. The report does not say whether that remainder represents people comfortable with conditions, people who did not answer, or another response option. Equally, the "resigned" and "would stop using" columns appear to be subsets of the discomfort total, but the document never states how the discomfort figure was constructed.

Germany at the top, the US and UK at the bottom

Country, according to Usercentrics, is the only variable that changes the overall picture. Chapter two of the report presents resigned consent and flat refusal for each of the seven markets:

Market Resigned consent Flat refusal
Germany 24% 16%
Spain 19% 25%
Italy 19% 21%
Netherlands 14% 31%
Sweden 14% 29%
United States 10% 25%
United Kingdom 11% 24%
Global average 16% 23%

Germany is the outlier on both measures. Its resigned consent rate of 24% holds between 23% and 24% in all six data categories, according to the report, so the average is not being pulled up by a single type of data. Germany also has the lowest flat refusal rate of any market, at 16% against a 23% global average, and, according to the report, the lowest refusal rate in every category tested. "Germany is the market most ready to say yes, and that yes is the most reluctant of any market in the study," Usercentrics writes.

The company offers an explanation drawn from the wider State of Digital Trust 2026 dataset: German consumers rank trust in the regulatory framework above brand transparency as their reason for granting access, which Usercentrics describes as the opposite of every other market studied. "That's plausibly why comfort there runs high but genuine enthusiasm doesn't: people are relying on the rules to catch problems, not on the request itself feeling right," the report states. The underlying ranking is not reproduced in the Resigned Consent Report, so the figures behind that claim cannot be checked from the document itself.

The United Kingdom and the United States sit at the other end, at 11% and 10%. The Netherlands and Sweden show the highest flat refusal rates, at 31% and 29%. According to the report's closing chapter, "Swedish and Dutch consumers are more likely to simply leave than resign themselves to giving data access." That sentence is worth reading against the chart, which plots flat refusal rather than abandonment for each market. The report does not publish country-level figures for the "would stop using the AI entirely" response.

German reluctance sits awkwardly with an earlier finding from the same survey. When PPC Land covered the main State of Digital Trust 2026 report on June 27, 2026, Usercentrics reported that 73% of German consumers would pay more for AI transparency - the highest share in the study - at an average premium of 9%, and that 75% of Germans had taken some action against a brand over AI data concerns. A market that pays for transparency, acts against brands and still says yes most reluctantly fits the company's "resigned" framing, though the report does not connect these figures.

Age changes how often people say yes, not how they mean it

Chapter three splits the data by generation. The figures, as read from the report's charts:

Generation Resigned consent (avg) Flat refusal (avg) Would stop using AI entirely (avg)
Gen Z (18-29) 17% 21% 14%
Millennials (30-44) 17% 20% 14%
Gen X (45-59) 16% 26% 18%
Boomers+ (60+) 13% 30% 27%

The break comes at 45. Below that age, Gen Z and Millennials behave almost identically. Above it, refusal climbs to 26% for Gen X and 30% for those aged 60 and over, while abandonment rises from 14% to 18% to 27%. Resigned consent moves the other way, falling to 13% among the oldest group. "Both age blocs are uncomfortable. They just resolve it differently: under-45s stay and resign, over-45s are more likely to refuse outright or leave," the report states.

The more interesting claim follows. According to Usercentrics, when the analysis looks only at the people in each generation who said yes, "roughly a quarter to a third of those are reluctant, in every generation, Gen Z included." The company's conclusion: "Age doesn't decide how reluctant a 'yes' is. It decides how often someone gets to yes in the first place." The report does not publish the base - the share of each generation that granted access at all - so the quarter-to-a-third range cannot be recalculated from the document.

Two attitude measures sit alongside the behavioural ones. The share strongly agreeing that they trust AI less than humans with their personal data runs from 23% for Gen Z and 22% for Millennials to 24% for Gen X and 30% for Boomers+. The share very concerned that AI personalization feels intrusive runs from 24% and 25% for the two younger groups to 30% for Gen X and 37% for the oldest. According to the report, the groups reporting the least unease are also the least likely to refuse outright, but "once they say yes, the reluctance behind it looks about the same as everyone else's."

Method, and what the report leaves out

The methodology is limited to a single box on the final page: "11,000 consumers. 7 markets. Conducted by Sapio Research in March 2026. Accurate to ±0.9% at 95% confidence. Sweden is a new market this year; YoY comparisons exclude Sweden." The September 29, 2026 press release distributed through Business Wire names the seven markets as the UK, the US, Germany, Spain, Italy, the Netherlands and Sweden, and describes this as the second annual edition of the study.

Several elements are missing. The report does not give the sample size in each country, which matters because the ±0.9% margin applies to the full 11,000 sample, not to individual markets or age groups. Subgroups of 1,000 to 2,500 respondents would carry margins several times wider. For comparison, the 2025 edition of the survey split its roughly 10,000 respondents as 2,500 each in the US and Germany, 2,000 in the UK and 1,000 each in Spain, Italy and the Netherlands. If a similar split applies this year, the Dutch and Swedish figures rest on far smaller samples than the German and American ones; the report does not say.

The exact wording of the questions is not published. Nor does the report explain how respondents were sampled or weighted, or whether the AI access scenario was presented as a hypothetical or tied to products respondents already use. The report mentions year-on-year comparisons excluding Sweden, yet contains no year-on-year figures.

The fieldwork date also deserves attention. The survey was conducted in March 2026 and the Resigned Consent Report was published at the end of September, six months later. The PDF file carries a creation date of September 23, 2026, though neither the report nor the landing page prints a publication date.

Discrepancies between the documents

The figures do not line up perfectly across Usercentrics' own materials.

  • 16% or 17%. The report and the landing page both give the average resigned consent rate as 16%. The September 29 press release, according to the version published on Usercentrics' own press page, gives 17%. Seventeen percent is the peak figure for three individual categories, not the average.
  • 7% or 8%. The Resigned Consent Report puts the share fully comfortable with no conditions at 7%. PPC Land's June 27 coverage of the parent report, drawn from the same March 2026 fieldwork, cited 8% for the same measure. The category table in the new report, with values of 6% and 8%, averages to roughly 7.3%, which may explain the rounding.
  • "Comfortable" and "uncomfortable." The web landing page labels the 16% figure as consumers "granting that access despite feeling comfortable," dropping the "un" that appears in the PDF. The PDF wording matches the rest of the report.
  • Access versus discomfort. The opening section states that "close to six in ten consumers say they would let an AI assistant access their data," then cites 58% as the average share reporting discomfort. The two six-in-ten figures describe different things, and the report does not publish an overall access rate that would support the first claim.
  • Two to one everywhere. The report says reluctant access outnumbers willing access two to one "across every category and every generation." The category table supports this, with the narrowest gap in company data (16% against 8%). The generation section, however, does not publish a "fully comfortable" figure for any age group, so that half of the claim cannot be verified from the document.
  • Charts on the web page. The landing page reproduces the country and generation charts without their data labels, so the percentages above are taken from the PDF.

A report built around a product catalogue

Usercentrics sells the tools its report recommends. Each of the four chapters ends with a "Take action" box pointing to a product: Cookiebot CMP after the category analysis, Usercentrics CMP after the country analysis, Usercentrics Server-Side Tracking after the generation analysis, and MCP Manager after the final chapter. The company is one of the largest vendors of the consent management platform software that produces the consent rates the report says cannot be trusted at face value. Usercentrics surpassed €100 million in annual recurring revenue in October 2025, processing more than 7 billion consents a month across more than 2 million websites and apps.

That commercial position cuts both ways. A consent vendor has an interest in persuading clients that their opt-in rates hide fragility, since the remedy it proposes - redesigned flows, granular choice, server-side data routing - is what it sells. The same vendor also has an unusually direct view of how banners perform. The report does not draw on that operational data. All of its figures come from stated survey responses, not from observed behaviour on client websites.

The pitch boxes describe the products in specific terms. According to the report, Cookiebot CMP "gives equal visual weight to accept, decline, and customize, and flags where a flow is nudging people toward the click-through." The server-side tracking product "routes data to your own backend before it reaches a third party." MCP Manager governs what a company's own AI agents can access when connected to company systems and customer data - a reference to the MCP protocol that AI agents use to connect to external tools. The report also lists GDPR, CCPA and CIPA among "20+ other frameworks" its CMP handles.

Four design checks

The final chapter sets out four principles for consent design, which mirror long-standing regulatory positions on banners. The report lists them as giving every option equal weight, letting people choose by category rather than all-or-nothing, making opt-out as easy as opt-in, and asking at the moment the access request becomes relevant rather than in advance.

The first and third principles describe the practices regulators have targeted as a dark pattern. France's CNIL ordered several website publishers in December 2024 to fix cookie banners that made rejection harder than acceptance, citing reject options styled as low-visibility links and accept buttons repeated several times. In Germany, the Hannover Administrative Court ruled in March 2025 that a publisher's banner "specifically guides users toward providing consent and prevents them from rejecting cookies," because acceptance took two clicks while rejection required several menus.

The report is candid about the commercial cost of following its own recommendations. "Consent experiences built this way are likely to generate lower opt-in rates than the one you have today. That's the point," it states. Usercentrics frames the objective in one line: "The goal isn't a higher yes rate. It's a more honest one."

Tilman Harmeling, who works in Strategy and Market Intelligence at Usercentrics, is the only named source in the document. "A resigned yes and a real one look identical in your consent rate, but behave nothing alike. One holds under pressure. The other holds until the first bad headline. If you can't tell which one you have, you don't know how much of your AI adoption is actually yours to keep," Harmeling said in the report.

For advertisers and publishers, a consent rate is not an abstract trust indicator. It determines how many users can be tracked, measured and targeted, and in many setups it decides how much conversion data must be modelled rather than observed. A report arguing that a portion of those consents is fragile goes to the core of how digital advertising measures itself in Europe.

The question of how often users are asked is also live in Brussels. The European Commission's Digital Omnibus proposal, published in November 2025, included Article 88b, which would have required websites to honour machine-readable consent signals from browsers. The Commission's explanatory memorandum, as PPC Land reported in June 2026, described a regulatory solution to "consent fatigue and proliferation of cookie banners" as long overdue. The Council deleted the browser signal from its text in June 2026, and a September 3, 2026 Presidency draft returned cookie rules to the ePrivacy Directive while barring a new request for the same purpose for at least six months after a refusal. On September 10, 2026, 19 organisations asked EU institutions to restore Article 88b, arguing that banners produce acceptance in up to 90% of cases while about 3% of people want tracking - figures the coalition did not tie to a study.

Usercentrics' consent-fatigue reading of its data lands in the middle of that debate. If repeated prompts wear down resistance without removing the underlying concern, as the report contends, then both a banner-heavy status quo and a rule that limits re-prompting carry consequences for the quality of the consents collected. The report does not take a position on the Omnibus.

The AI angle connects to a run of consumer surveys PPC Land has tracked through 2026. Shift Browser's US survey, published in March, found 81% of 1,448 respondents concerned about AI systems accessing personal data, while 32% used AI daily. A Cloaked survey of 1,009 US adults, covered in May, found that 18% trust AI platforms to keep their data secure and 87% are uncomfortable sharing financial information with AI - consistent with the financial-accounts result in the Usercentrics data. Usercentrics itself partnered on an MIT Technology Review Insights report in April 2026 arguing that privacy-led design had become a marketing priority.

What none of these studies do is measure what happens after a reluctant yes. The Resigned Consent Report asserts that "genuine consent shows up as loyalty and retention" while "resigned consent shows up as churn and complaints," but offers no longitudinal data to show it. The claim is a hypothesis about future behaviour drawn from a single round of stated preferences. The parent report's finding that 24% of consumers had cancelled a subscription or stopped buying over AI data concerns is the closest the survey comes to behavioural evidence, and it does not distinguish between people who had previously consented reluctantly and those who had not.

Timeline

Summary

Who: Usercentrics, a Munich-based consent management company that sells Cookiebot CMP, Usercentrics CMP, server-side tracking and MCP Manager, with fieldwork by Sapio Research. Tilman Harmeling of Usercentrics' Strategy and Market Intelligence team is the only named source.

What: The Resigned Consent Report, a 12-page analysis of the State of Digital Trust 2026 survey, finds that 7% of consumers are fully comfortable granting AI access to their data without conditions while 16% grant it despite discomfort. Germany has the highest resigned consent rate at 24%, against 11% in the UK and 10% in the US. The press release gives the average as 17%, and the report does not publish question wording or per-country samples.

When: Published September 29, 2026, from fieldwork conducted in March 2026.

Where: Seven markets: Germany, Spain, Italy, the Netherlands, Sweden, the United Kingdom and the United States.

Why: The report argues that consent rates conflate genuine and reluctant agreement, which matters for advertisers and publishers who rely on consent to track and measure users, and for EU policymakers weighing how often websites may ask for consent under the Digital Omnibus.