Facial recognition software startup Clearview AI has been under fire over revelations it’s been training its machine learning algorithm on billions of publicly available photos, including public personal images scraped from social media without permission.

Now Facebook has become the latest tech giant to tell the controversial company to quit it.

A spokesperson for Facebook told Buzzfeed News that it has sent “multiple letters” to Clearview asking that it cease and desist scraping “data, images and media” from Facebook itself and from Instagram (which it also owns) and restating its policies as well as asking for further details about their work and activities.

YouTube and Venmo also say they have sent or are sending cease and desist letters to Clearview.

Twitter told the company back on Jan. 21, according to a New York Times report, to stop scraping images on the platform, which is a violation of their terms of service, and to delete any data collected so far. 

Even if Clearview complies with all deletion requests, the question remains of how much of that scraped data remains in some way in their systems — if, even once deleted from the database, the AI can still use what it’s learned from those images in future applications and searches. 

Clearview claims it’s scraped over 3 billion publicly available images from websites and social platforms, using sources ranging from company websites to faces in the blurry background of gym selfies, and can now identify just about anyone.

Clearview has been aggressively marketing to law enforcement and government, encouraging police to test the software in investigations, licensing it to police departments, and boasting in corporate presentations of their proposed expansions to 22 more countries including the United Arab Emirates, Singapore, Qatar, and Australia. It has also stated it plans to create AR glasses and other wearables. 

The software’s use by police was banned in New Jersey by the state’s attorney-general last month. AG Gurbir Grewal called it “chilling.”

The company’s CEO, Hoan Ton-That, told CBS News in a sit-down interview that the company’s approach is actually a free-speech issue, with their right to “public information” protected by the First Amendment. Their overall argument is basically that you put it on the internet, it’s fair game — even for a company that’s then going to sell your data, or software based on it, to law enforcement and authoritarian regimes.

And while the company insists its software is “not a surveillance tool,” it really looks, walks, and quacks like one. Hell, when Facebook thinks you’re taking it too far, that’s really cause for a little self-reflection.