One primary non-scam feature is to allow a technician to remote in to a desktop and use authentication that should not be viewed by the user. I actually worked for an ISP where a tech actually telnetted into a Cisco router that was misbehaving via my computer remotely, and they didn't blank the screen.
Back then, you could have seen the domain name, username, and password. Had I wanted to, I could have captured this information and used it to cripple large portions of the Internet, and it would be virtually untraceable back to me, and the technician who leaked their credentials would have taken the fall.
The same goes for remote workers. In some versions and types of Remote Desktop, the desktop is visible locally. This means that anyone casually sitting at your computer might see you type in usernames, passwords, credit card numbers, and any other PII that you might wish to have hidden from the general public.
Like many tools we use in real life (knives, guns, lock picks, etc), black screens have intended and unintended uses. Black screens protect data leaks and hacks for sensitive systems, and should definitely be used if you're remoting into a remote device and need to protect some data from casual observation.
It's unclear that there's a clear fix for this, because even if you add a "are you sure you wish to allow the remote user to blank your screen and scam you" prompt, people who are gullible enough to install such apps are just as likely to trust the "technician" on the other end of the line that says "Oh, don't worry about that, just click OK so I can <whatever-scam-here>".
I agree that inherently, this feature clearly will continue to be abused, but there are legitimate reasons why you'd want to have this feature available. Not having this available would lead to many more serious data leaks and havoc that hackers could use to bring down entire infrastructures.