Pictures is just a pretense. Step #2 will be URLs of pictures that users send each other using the phone's keyboard. Step #3 will quietly expand the definition of a URL to any keyboard input. At that point, Apple will have a god's eye view into all communications of their users, be it over WhatsApp, Signal, SMS or an online forum. Imagine having a dashboard that shows in realtime with gps coordinates how many text messages contain the word "enough"! The same scanner module can detect certain words in phone calls. What dictator wouldn't want that?