Majority of Americans trust what’s online less than ever

The average American trusts less than half of what they see and read online, according to new research.

A survey of 2,000 U.S. adults found that skepticism is at an all-time high when it comes to the internet.

According to the results, Americans believe only 41% of what they consume online is totally accurate, fact-based and created by a real human.

In fact, they also believe that 23% is completely false and purposely inaccurate or misleading, while 36% falls somewhere in between.

Three-quarters say they trust the internet less today than ever before.

Another 78% agree that the internet has “never been worse” when it comes to differentiating between what’s real and what’s artificial.

To put things into perspective, the average American comes across information that they know or suspect was generated by AI about five times per week, with 15% indicating it's more than 10 times.

Social media posts (48%), news articles (34%) and chatbots (32%) are the top three most likely culprits when it comes to AI-generated or misleading content.

In fact, those polled believe that 50% of the news stories and articles they come across online have some element of AI, whether it be images or the actual written content.

Despite their suspicions, less than a third (31%) are confident in their ability to differentiate if a product or service review was written by an AI or an actual human being.

The survey, conducted by Talker Research on behalf of World, even went so far as to test Americans on their ability to do just that.

When shown a set of business reviews written by AI and humans, only three in 10 (30%) could correctly identify which was which.

Of the three options that were written by people, two ranked at the very bottom of the list, demonstrating just how easy it is to be duped in this day and age.

With 80% of Americans relying on reviews in some capacity when it comes to choosing a business to support, it’s no surprise that bot-written reviews (62%), AI customer service reps (50%) and AI-generated images (49%) make them less likely to patronize that company.

Unfortunately, 46% of those polled have actually purchased something that ended up not being what was advertised, and 24% of those weren’t able to get a refund or return the item.

"Trust in the internet hasn't just declined — it's collapsed under an avalanche of AI-generated noise. The internet has become a house of mirrors where 78% of Americans can no longer distinguish real from artificial,” said Rebecca Hahn, Chief Communications Officer of Tools for Humanity, developers of World ID. We offer an elegant solution: instant human verification without compromising privacy. No personal data, no tracking — just a simple proof that restores what's been lost: trust in who we're actually connecting with online."

The survey also revealed that the most stressful situation when it comes to differentiating whether they’re dealing with a person or chatbot is when speaking to a customer service representative (43%).

Booking lodging or hotels (23%) and sending money through a third-party app (22%) also ranked in the top three most anxiety-inducing situations.

In order to verify human status, Americans have a few tricks up their sleeve. One respondent said, “I often ask open-ended questions or test for human-like responses, such as asking for personal opinions or experiences.”

Another asks questions, including: “What's your birthday? What's your mother’s maiden name?  Where were you born?  Do you have any distinguishable marks on your skin?”

Beyond that, 24% will Google or search for the entity online to verify their human status, while 23% ask for a phone or video call.

For all their struggles, 82% of Americans agree that businesses and vendors should be legally required to disclose whether AI is used in their marketing, content, customer service or on their website.

"Being able to prove you're human online is becoming as essential as having an email address was twenty years ago,” said Hahn. “Our survey shows Americans are desperate for tools that restore confidence in digital interactions. We’re pioneering a new paradigm where human verification becomes a foundational layer of the internet — simple, secure, and universally accessible. This isn't just about solving today's trust crisis; it's about building tomorrow's internet where human-to-human connection remains at the heart of everything we do."

Survey methodology:

Talker Research surveyed 2,000 general population Americans; the survey was commissioned by World and administered and conducted online by Talker Research between March 28 through March 31, 2025.

We are sourcing from a non-probability frame and the two main sources we use are:

• Traditional online access panels — where respondents opt-in to take part in online market research for an incentive
• Programmatic — where respondents are online and are given the option to take part in a survey to receive a virtual incentive usually related to the online activity they are engaging in

Those who did not fit the specified sample were terminated from the survey. As the survey is fielded, dynamic online sampling is used, adjusting targeting to achieve the quotas specified as part of the sampling plan.

Regardless of which sources a respondent came from, they were directed to an Online Survey, where the survey was conducted in English; a link to the questionnaire can be shared upon request. Respondents were awarded points for completing the survey. These points have a small cash-equivalent monetary value.

Cells are only reported on for analysis if they have a minimum of 80 respondents, and statistical significance is calculated at the 95% level. Data is not weighted, but quotas and other parameters are put in place to reach the desired sample.

Interviews are excluded from the final analysis if they failed quality-checking measures. This includes:

• Speeders: Respondents who complete the survey in a time that is quicker than one-third of the median length of interview are disqualified as speeders
• Open ends: All verbatim responses (full open-ended questions as well as other please specify options) are checked for inappropriate or irrelevant text
• Bots: Captcha is enabled on surveys, which allows the research team to identify and disqualify bots
• Duplicates: Survey software has “deduping” based on digital fingerprinting, which ensures nobody is allowed to take the survey more than once

It is worth noting that this survey was only available to individuals with internet access, and the results may not be generalizable to those without internet access.