Flock Uses Overseas Gig Workers to Build Its Surveillance AI

Flock, the automatic a license plate reader and AI-powered camera company, uses foreign workers from Upwork to train its machine learning algorithms, with training materials instructing workers how to review and categorize images, including images of people and vehicles in the United States, according to material reviewed by 404 Media that was accidentally exposed by the company.

The findings raise questions about who exactly has access to the footage collected by Flock surveillance cameras and where the people reviewing the footage may be based. Flock has become a ubiquitous technology in the United States, with its cameras present in thousands of communities that cops use daily to investigate cases such as carjackings. Local police also conducted numerous ICE searches in the system.

Companies that use AI or machine learning regularly turn to foreign workers to train their algorithms, often because labor is cheaper than hiring domestically. But the nature of Flock’s business – creating a surveillance system that constantly monitors the movements of US residents – means the footage could be more sensitive than other AI training tasks.

Flock’s cameras constantly scan the license plate, color, make and model of all passing vehicles. Law enforcement is then able to search cameras across the country to see where a vehicle has driven. Authorities typically search this data without a warrant, leading the American Civil Liberties Union and the Electronic Frontier Foundation to recently sue a city covered by nearly 500 Flock cameras.

Broadly speaking, Flock uses AI or machine learning to automatically detect license plates, vehicles and people, including the clothes they are wearing, from camera images. A Flock patent also mentions “race” detecting cameras.

Several tipsters directed 404 Media to an exposed online panel that showed various metrics associated with Flock’s AI training.

It included figures for “annotations completed” and “annotator tasks remaining in queue,” with annotations being the notes that workers add to reviewed images to help train AI algorithms. Tasks include categorizing vehicle makes, colors and types, transcribing license plates and “audio tasks”. Flock recently began promoting a feature that will detect “screams.” The panel showed that workers sometimes completed thousands and thousands of annotations over two-day periods.

The panel on display included a list of people responsible for annotating Flock’s images. Taking these names, 404 Media discovered that some were in the Philippines, according to their LinkedIn and other online profiles.

Many of these people were employed by Upwork, according to the exposed documents. Upwork is a freelance work platform where businesses can hire designers and writers or pay for “AI services,” according to the Upwork website.

Informants also pointed to several publicly available Flock presentations that explained in more detail how workers should categorize images. It’s unclear what specific camera footage Flock’s AI workers are looking at. But screenshots included in the workers’ guides show numerous images of vehicles with U.S. plates, including in New York, Michigan, Florida, New Jersey and California. Other images include road signs clearly showing that the images are taken from the United States, and one image contains an advertisement for a specific law firm in Atlanta.