Australian consumers are among the least willing globally to allow artificial intelligence (AI) to make decisions independently, according to new research from Ping Identity, raising questions about trust, security and oversight as businesses introduce more autonomous AI agents.
The company’s 2026 Consumer Survey, based on responses from 11,000 consumers across 12 countries, found that just 18% of Australians were willing to let AI make independent decisions, compared with 70% in India and 61% in Indonesia.
Globally, consumers showed a clear preference for AI assistance over autonomous action. While 56% were comfortable using AI for low-stakes support or recommendations, only 5% were comfortable with AI taking action subject to their approval, and just 3% supported AI performing routine transactions without final approval.
The findings come as organisations increasingly explore agentic AI systems capable of completing tasks, accessing information and making decisions on behalf of users.
However, granting AI agents greater autonomy introduces questions about how access to personal information and digital accounts should be managed.
The research found that just 12% of consumers fully trusted organisations to manage their identity data, down from 17% in 2025. Financial access was a particular concern, with 42% of respondents wanting to limit an AI agent’s spending to no more than $100, while 22% would prohibit AI agents from making financial transactions altogether.
Consumers were also concerned about how AI agents might represent them, with 71% worried that agents acting on their behalf could misrepresent their personality or values. Nearly half (47%) considered the need to continually review an AI agent’s work a greater concern than the possibility of the technology making a mistake.
Darryl Jones, Vice President of Consumer Segment Strategy at Ping Identity, said businesses needed to establish clear boundaries as AI systems assumed greater responsibility.
“Consumers are open to AI doing more, but they want a say in how far it can go,” Jones said. “The more responsibility we give AI, the more important it becomes to be clear about what it can do, what it can access and who remains in control.”
Transparency was another priority, particularly in Australia, where 68% of consumers wanted to know whenever AI was being used, compared with 63% in the US and 57% in the UK.
Globally, 56% of respondents considered knowing when AI was being used more important than receiving the fastest possible result.
The findings suggest organisations deploying agentic AI will need to consider how permissions, identity controls and human oversight are built into their systems, particularly when AI agents are given access to sensitive information or authority to act on behalf of customers.
