There's a mild concern here, but it does rely on two things - having a known image and having it uploaded to a web service. Judging by the information in the article an original image wouldn't be matched, and it is in the ToS (good luck reading all of that).

The organised groups that make money off this are much wiser and harder to catch.

More interesting is the quote "Microsoft also developed Photo DNA to match pictures using just a pixel". I'm not sure it works quite how they suggest.