Consumers are most comfortable with AI providing support and recommendations. Just 5% say the highest level of responsibility they would give AI is taking action with their approval, while 3% would allow routine actions or transactions without final approval.

Nearly one-third (31%) are willing to let AI independently make certain lower-stakes decisions on their behalf, such as scheduling appointments, selecting products, or exploring rentals or homes. This signals meaningful openness to AI delegation, even as consumers remain cautious about giving AI greater responsibility.

57% of consumers trust AI to some degree, and 43% say they trust AI more today than they did one year ago.

71% are concerned that an AI agent acting on their behalf could misrepresent their personality, ethics or personal values. 47% are more concerned about constantly having to double-check an agent’s work than about it making a mistake.

Yes. 56% say knowing whenever AI is being used matters more than simply getting the fastest possible result.

Consumers would trust an AI agent with an average of US$200 for routine purchases. 42% would limit AI to $100 or less, while 22% would not trust AI with any amount of money.

When consumers suspect a website is using undisclosed AI, 20% say they would leave the website immediately, and 20% would avoid using that brand or company in the future.

As AI agents begin acting on behalf of people, consumers want confidence in who is acting, what authority AI has and how those actions remain accountable.