AI agents pose financial risks without oversight, warns Dr Karen Sutherland from UniSC on shopping impacts
AI shopping agents could make online purchases more convenient, but giving them access to payment details, emails and personal information raises questions about financial security, privacy and consumer control.
On Social Intelligence with Dr Karen Sutherland, Professor Karen Freberg from the University of Louisville in Kentucky discusses the risks of allowing artificial intelligence to make purchasing decisions on behalf of consumers.
As AI agents become increasingly capable of researching products, comparing prices and completing transactions, consumers face a new challenge: deciding how much autonomy these tools should have over their money and personal data.
How much control should AI have?
AI shopping agents promise to simplify everyday tasks by making purchasing decisions and managing transactions with minimal human involvement. However, granting these systems broad access to financial accounts and personal communications could introduce new risks.
Consumers may need to consider whether an AI agent should be able to make purchases independently, access stored payment information or use details from private emails to inform its decisions.
Without clear boundaries, automated purchasing could result in unintended transactions, inappropriate spending or personal information being shared without proper authorisation.
The risks of handing over personal data
Allowing AI agents to access emails and payment details creates additional privacy and security concerns.
Information gathered for one purpose could potentially be used in ways consumers did not anticipate, particularly when permissions are broad or difficult to manage.
The issue extends beyond whether an AI agent can complete a purchase correctly.
Consumers also need to understand what information the system can access, how that information is handled and whether it can initiate transactions without further approval.
For businesses developing these technologies, building consumer trust will depend on transparency, clear permissions and effective safeguards.
Setting boundaries before AI spends
A central question for consumers is which decisions should remain under human control.
AI agents may be useful for comparing products, identifying deals and recommending purchases.
However, spending limits, approval requirements and restrictions on access to sensitive information could help reduce the risk of unwanted transactions.
Consumers should understand what they are authorising before connecting an AI agent to their accounts, particularly when the system can access payment methods or act on their behalf.
The discussion with Professor Freberg highlights a broader challenge as AI moves from providing information to taking action. The convenience of automated shopping must be weighed against the need to protect personal data, financial security and individual choice.
As these tools become more integrated into everyday life, establishing clear rules around consent and accountability will be critical to ensuring consumers remain in control of their purchases.
Watch Social Intelligence with Dr Karen Sutherland for more on the opportunities and challenges emerging from artificial intelligence.
