I would also like to Azure AD Identity Protection service start considering or flagging if an Unknown IP, Unfamiliar Location, or Imposible traveller risks are being impacted by a Anonymising VPN Service, especially for publicly accessible common services (Exchange Online, SharePoint/OneDrive, etc). I'm aware that not all of the providers would be publishing their server IP addresses, or that they would be using IP addresses under their own ASN, or that any list would be complete or up to date.
That said, if there was just a flag next to the IP address identifying it as a 'Possible Public VPN Traffic' or something like that and still leaving the decision up about how to act with the person reviewing the alert. Of course if it was in the logic you could configure for conditional access or even the detection algorithm then all the better.
I'm also aware that this would not be an easy problem to solve, but it's a trend i'm seeing in false positive alerts and even if you mandate via policy not to use them or only use Corporate VPN's, sometimes the alternatives (Hotel Internet, Free Wifi, etc) may be worse, especially if its a personal device or you are a cloud only style of user/organisation.