What is the World’s “Trust Tipping Point” for Artificial Intelligence?
It appears that a significant portion of people—66% to be exact—are quite comfortable with the idea of using AI in both their personal and professional lives. Yet, they seem to have certain boundaries regarding how far they’re willing to rely on it.
A recent global study involving 11,000 participants aimed to understand the extent to which AI is affecting various facets of daily life. The aim was to see how this integration impacts trust, current utilization, and future expectations regarding AI.
The findings suggest that the average person consciously engages with AI roughly six times a week, with individuals in India using it about 14 times weekly, and those in Indonesia averaging just over 10 times.
Interestingly, respondents believe that they unintentionally interact with AI almost four times a week and suspect they encounter it at least that often too.
When it comes to interacting with brands, a majority—56%—express a preference for being made aware of when AI is in use, which is considerably higher than the 26% who prioritize obtaining fast, accurate results, even if AI’s involvement isn’t disclosed.
The absence of disclosure regarding AI seems to raise red flags; 39% indicate increased caution about sharing personal information, while 20% would avoid the brand altogether or leave the website if they suspected undisclosed AI usage.
Notably, respondents from the U.S. (23%) and Australia (24%) were the most likely to exit websites promptly upon suspecting they were dealing with undisclosed AI.
The survey, carried out by Talker Research for Ping Identity, sought to gauge consumers’ comfort levels with various aspects of AI, including the kinds of tasks they’d be willing to delegate to it and their trust concerning financial matters.
According to the results, people feel most at ease using AI for tasks related to general learning—like researching hobbies or summarizing literature (37%)—as well as for DIY assistance (32%) and creative projects like drafting emails or making fun images (32%).
Many respondents are also open to utilizing AI for work-related tasks such as summarizing meetings or writing reports (24%) and even for job-seeking activities like resume optimization or interview preparation (24%).
Additionally, AI might be trusted for reviewing cover letters (24%), planning meals (23%), researching potential new locations to live (22%), providing skincare suggestions (18%), or even reviewing job offers (18%).
While about 17% would trust AI with minor health issues, such as cuts or scrapes, that number drops for more complex areas like mental health guidance (12%) or medication recommendations (10%).
Hypothetically speaking, respondents would be willing to entrust an AI agent with around $200 for routine purchases without needing to approve each one.
However, only 31% actually feel comfortable allowing AI to make decisions for them—like scheduling or product selection—independently. Interestingly, 70% of participants in India are more inclined to trust AI for decision-making, which is the highest percentage among the surveyed nations.
Darryl Jones from Ping Identity noted, “Only 12% of those surveyed are ‘very aware’ of what agentic AI entails. Moreover, 70% express concerns that an AI agent might misrepresent their values, indicating a gap between the technology’s capabilities and consumer comfort levels.” He emphasized the importance of users knowing when AI is making decisions on their behalf.
Overall, a majority (57%) of those surveyed trust AI to some degree, with 43% noting that their trust has grown in the past year.
As trust increases, so do expectations. The data reveals that people believe ethical standards applicable to human sources should extend to AI as well. They want assurance that personal data remains confidential, that there’s an explanation for conclusions drawn, and that the AI is authentic.
Users generally prefer it when AI presents facts neutrally (38%), rather than confirming their pre-existing beliefs (21%) or providing differing viewpoints (20%).
Despite varying trust levels, 40% recall instances where AI has generated false information or “hallucinated” at least once. Interestingly, they tend to blame the companies behind the AI more than the users who elicited its responses (29% versus 15%).
Nonetheless, 44% consider these inaccuracies as unintentional mistakes instead of deliberate deceit (23%).
Jones remarked, “While it’s common to hold humans accountable for mistakes, assessing AI complicates matters since it lacks a human touch. Most respondents view AI primarily as a ‘search engine’ (48%), emphasizing the need for accuracy. It’s critical for users to verify information and be cautious about acting solely on AI’s recommendations without understanding their origins.”
Research Methodology
Talker Research surveyed 11,000 global respondents with internet access (2,000 each from the U.S. and U.K., and 1,000 each from France, Germany, Singapore, Australia, India, Indonesia, the Netherlands, Sweden, Spain, and Italy). The survey, commissioned by Ping Identity, was conducted online between June 17 and June 25, 2026.
