Merchant fraud prevention leaders are being urged to strengthen oversight and governance around AI shopping assistants as concerns grow over unauthorised transactions, liability disputes and consumer trust.
New research from Exploding Topics by Semrush found that while 78% of consumers have used AI to support shopping decisions in the past six months, only one in three would trust AI to actually complete purchases on their behalf.
The findings come as major retailers continue integrating AI-powered shopping tools into customer journeys. Amazon’s Rufus and Walmart’s Sparky assistants are already reshaping product discovery and purchasing behaviour, but the increasing autonomy of these systems is raising fresh concerns around fraud risk and accountability.
For merchants, the challenge extends beyond customer experience into operational and reputational exposure. Recent changes to Target’s terms and conditions reportedly place financial responsibility for AI-generated purchasing errors onto customers, while broader industry debate continues around how much control AI systems should have within payment flows.
Experts warn that the greatest risk lies in the transition from AI assistance to AI-led automation without adequate safeguards.
Among the key concerns identified are AI assistants completing purchases without explicit customer approval, insufficient transparency around recommendations and weak escalation pathways when errors occur.
For fraud prevention teams, this creates new complexities around transaction authentication, dispute management and customer trust. Over-delegating purchasing authority to AI systems could increase the risk of unauthorised transactions, biased recommendations or security breaches—particularly where human oversight is limited.
The research also highlights the importance of transparency around terms and conditions, data usage and recommendation logic. Consumers increasingly expect retailers to explain why products are being suggested and how personal data is being used to shape AI-driven decisions.
From a fraud and compliance perspective, AI shopping agents also introduce new governance challenges around liability attribution, authentication controls and behavioural monitoring.
Industry specialists say retailers will need to treat AI shopping systems as high-risk customer interaction points requiring robust oversight, auditability and clear accountability frameworks.
