How is any kind of antivirus or threat detection software supposed to operate on this standard?

Libel suits can be financially catastrophic, so even a tiny false positive rate could present risk that disincentivizes producing such software at all.

And a threat detection mechanism that has a 0.0% false positive rate is conservative to the point of being nearly useless.

(IAAL but this is not legal advice.)

It’s not libel. Defamation requires a false statement of fact. Marking a website as “unsafe” is an opinion.

They should be held legally culpable for libellous claims they make.

I dont care if their pre-LLM ai says "thingy bad". They are responsible for the scripts or black boxes they control. I dont care if they dont give a reason.

Claiming bad/malicious/etc site is 100% libel. And doubly so, anybody who has been forced to agree to a ToS with binding arbitration should have it removed for libel.