Consumers Embrace AI While Trust Lags

The market is moving toward visible, delegated automation as consumers accept more AI-driven actions. The EY global survey of 18,000 people finds 16% have used AI that acts on their behalf in the prior six months; 11% allow automatic cart refills and purchases, 11% give AI access to carry out financial tasks, and 9% have used self-driving vehicles. The report frames this shift as the rise of hyper-autonomy, where AI agents interact with each other to complete customer outcomes and decision authority migrates from humans to systems.

Designers and engineers must treat unattended AI as a distributed, transactional system with explicit controls for safety, traceability, and recoverability. Key engineering and governance constructs to adopt include:

• •Fine-grained authorization and consent to limit what agents can do on a user's behalf
• •Decision and audit logging to enable post-hoc review and dispute resolution
• •Human-in-the-loop gates and mechanisms for rollback or remediation for high-risk actions
• •Ongoing model monitoring, feedback loops, and explainability measures to support oversight

These controls must be integrated into product flows, APIs, and customer UX so override and remediation are obvious and low-friction.

Convenience is the proximate driver: users already accept AI for route planning, recommendations, and basic automation because outcomes are easy to inspect and correct. As EY notes, "By taking care of small, everyday tasks, AI has slipped into daily routines with little resistance." The risk profile changes as tasks become consequential, moving from observable, reversible recommendations to opaque, autonomous actions. That shift elevates product design, logging, privacy-preserving telemetry, and legal-compliance workstreams. Firms that only optimize for short-term convenience risk regulatory and reputational costs when autonomy causes harm.

Product teams need to instrument delegated actions for auditability and consent; regulators and standards bodies will increasingly demand demonstrable controls. Expect real-world pilots to drive both technical best practices and regulatory scrutiny.