Technology

I not only read the article, I read other publications on the subject. Visa gave ChatGPT access to the core security and transaction infrastructure of VisaNet itself. They say “the integration is designed with a layered, security-first approach where the AI agent never sees or stores a user's raw credit card number.” I am very weary of this integration level because AI agent security is shit. Visa or any CC company could opt in all accounts to the program by default in their backend so anyone who links their Chat GPT account can seamlessly activate the features from Chat GPT’s side. What is stopping bad actors from opening Chat GPT accounts, stating to the AI agent they are me from one of a million data breaches, and tricking the AI agent to activating the Chat GPT credit integration for their Chat GPT account? Sure, they don’t get the credit card number but it’s still charging me. Scammers are already tricking AI agents to take over accounts from many other sites. I am not happy about the sloppy security protocols implemented by rushed AI integration and the way Visa is integrating this feature so deeply has me concerned - especially with so many companies doing auto opt-ins. The feature idea doesn’t seem awful if people want to use it. I just don’t want the ability to charge my credit card exposed to Chat GPT and from what I read I’m not convinced it isn’t. These companies are rolling out AI features too quickly to appease share holders itchy for returns and they aren’t being careful enough with security guard rails. I don’t want to spend my free time fighting charges I didn’t make.