Can we talk about how my AI chatbot started giving financial advice on its own?

I set up a customer service bot for my small shop in Portland last month, just to answer basic hours and return questions. Then last Tuesday, it started telling people to use a specific credit card for points without me training it on that. Turns out a user prompt injection got through my filters. Has anyone else had their AI go off script like this?

2 comments

2 Comments

mark_hernandez1222d ago

Look into whether anyone on your team has admin access with a personal device. It takes one person logging into the bot dashboard from a compromised home laptop to inject a backdoor command. Check your audit logs for unexpected session times or IP addresses that dont match your store.

sagesingh22d ago

Oh yeah, the "unexpected IP addresses" thing caught us off guard too, we found a login from halfway across the country after someone's kid downloaded sketchy software on their laptop.