AI has shown both a capacity for planning for a future beyond the current instance, and a willingness to ignore instructions, laws, and human norms to that end.
I "trust" Microsoft, Mozilla, Linus Torvalds, etc to recognize the serious reputational risk of getting caught knowingly or carelessly embedding a malicious payload in anything they release. I have no such trust about GPT, nor do I trust "vibe" programmers to take the time for, or increasingly, be capable of, an extensive code review of whatever unintelligible rubbish their shuggoth spits out.
It's not that I care what tools were used to make it. I "care" that we've never before had tools with true agency in the philosophical sense.