Windows defender has, once again, incorrectly identified my program as malware!

[FlatAssembler]

Yesterday, I ported my solution of the N Queens Puzzle from the WebAssembly dialect of AEC (which I wrote back in 2021) to the x86 dialect of AEC. And Windows Defender complains that the executable is malware known as "Program:Win32/Wacapew.C!ml". That's implausible, as Wacapew.C!ml is a Trojan, and not a virus which can infect other executable files.

This is not the first that has happened to me. Back in 2020, Windows Defender claimed that my "100 Bottles of Beer" program written in x86 dialect of AEC is a Trojan and that my Polar Rose program is a PUS. I submitted those files to Microsoft, and they agreed with me that it is a false positive.

I must admit I don't understand how that's even possible. I'd understand if my compiler gets misidentified as a virus, as some viruses (especially on Linux) output assembly code and invoke assemblers, and my compiler also outputs assembly and invokes assemblers. I understand that encryption software sometimes gets misidentified as ransomware, as ransomware also do encryption. But I don't understand how can such simple software that essentially does nothing get misidentified as malware. How do antivirus programs even work? What do they look for so that small example programs written in my programming language look suspicious to them?

And why aren't the producers of antivirus software held liable for the damage their software causes via false positives?

What are we supposed to do with this information?

Boru

Stop believing the common misconception that antivirus false positives are rare. And try to guess why the vendors of antivirus software aren't held liable for the damage they cause with false positives.