Your argument is essentially the same as the argument that the pirate bay didn't infringe copyright, it only facilitated infringement.

And we all saw how well that went legally.

Actually pirate bay was even less of an infringement as they did not dsitribute the copygihted content or derivatives themselves, only indexed where it could be found. With Copilot all the content you're getting goes trough Microsoft.

Machine learning is not important to this line of argument. We are talking about the legal responsibility of a tool.

Pirate Bay couldn't be used to do anything but infringe copyright, practically. That's not true for Copilot.

Nonsense. It tracked millions of legitimate torrents.