In Brief: A global survey of 11,000 respondents found that people knowingly use artificial intelligence about six times a week and encounter it unintentionally about four more times. Although most respondents expressed some trust in AI, they wanted companies to disclose its use, protect personal data and explain how automated recommendations or decisions were reached.
-
Image Credit Ping Identity
AI Becomes Part of Everyday Life
A global survey of 11,000 respondents, conducted by Talker Research on behalf of Ping Identity, found that people knowingly use artificial intelligence an average of 5.6 times per week. Participants estimated that they unintentionally interact with AI another 4.3 times per week and suspected they encounter it an additional 4.3 times.
Usage varied considerably by country. Respondents in India reported knowingly using AI about 14.2 times per week, or about twice a day. Those in Indonesia reported an average of 10.5 interactions per week.
Overall, 66% of respondents said they were comfortable using AI in their personal and professional lives. Acceptance depended partly on whether people knew when the technology was involved and retained control over important decisions.
Talker Research conducted the survey on behalf of Ping Identity. It examined AI use, trust, disclosure expectations and attitudes toward systems that can act independently for users.
Disclosure Influences Consumer Trust
When dealing with a company, 56% of respondents said knowing whenever AI was being used was the most important consideration. By comparison, 26% prioritized the fastest, most accurate result, regardless of whether the company disclosed AI involvement.
A lack of disclosure affected how respondents said they would behave. Thirty-nine percent would become more cautious about sharing personal information if they suspected a company was using undisclosed AI. Twenty percent would avoid that company in the future, while the same percentage would immediately leave its website.
Respondents in Australia and the United States were the most likely to leave a website over suspected undisclosed AI use, at 24% and 23%, respectively.
Consumers Favor Lower-Risk AI Tasks
Respondents were most comfortable allowing AI to help with general learning, including researching a hobby or summarizing a book. 37% selected that category. DIY assistance and creative tasks, such as drafting an email or generating an image for personal use, each accounted for 32%.
About 24% were comfortable using AI to summarize meetings, write reports, improve a resume, prepare for an interview or review a cover letter. Other accepted uses included planning meals, researching a possible place to live, providing skin care suggestions and reviewing an offer letter.
Acceptance declined for health-related tasks. Seventeen percent would use AI for assistance with a minor cut or scrape, and 16% would use it to better understand a chronic medical condition. Only 12% were comfortable receiving mental health advice from AI, while 10% would accept prescription medication recommendations.
Limits Remain on Independent Decisions
On average, respondents said they would hypothetically let an AI agent spend $200 on a routine purchase without needing final approval. Nevertheless, only 31% said they were likely to trust AI to independently schedule appointments, select products or explore rental options.
India recorded a substantially higher level of acceptance, with 70% of respondents there saying they were likely to trust AI to make decisions on their behalf.
Awareness of agentic AI remained limited. Only 12% said they were very aware of the concept. Seventy percent were concerned that an AI agent could misrepresent their personality or values when interacting with others.
Trust Rises Alongside Expectations
Fifty-seven percent of respondents said they trusted AI to some degree, and 43% reported trusting it more than they had one year earlier. At the same time, respondents expected AI systems to meet standards similar to those applied to human sources.
The three leading expectations were that personal information remain confidential, that the source explain how it reached a conclusion or recommendation, and that it provide evidence of its identity.
Respondents also preferred neutral information. Thirty-eight percent said AI was most useful when it presented facts without adopting a position. That compared with 21% who preferred confirmation of their existing views and 20% who valued a contrasting perspective.
False Information Continues to Affect Confidence
Forty percent of respondents said they had encountered AI-generated information that was false or fabricated at least once. Twenty-nine percent placed responsibility for such mistakes on the company that developed the AI, compared with 15% who blamed the person who entered the prompt.
Most did not view inaccurate output as deliberate deception. Forty-four percent described AI hallucinations as unintentional mistakes, while 23% considered them intentional.
Nearly half of respondents (48%) viewed AI primarily as a search engine or information source.