It’s difficult to see how artificial intelligence systems work, and to see whose interests they work for. Regulation could make AI more trustworthy. Until then, user beware.
This is like asking "can you trust a wrench?" It's a tool. You can trust it to operate within its parameters. The problem is that most of us don't understand what those parameters are; it's a black box. So yeah, if you're gonna use it, use it for low stakes applications are prepare to exercise considerable oversight ... like a harbor freight table saw.