Flock uses outside workers to build its AI surveillance system

flock, automatic The license plate reader and AI camera company uses outsourced staff from Upwork to train its machine learning algorithms, with training materials telling workers how to review and grade footage including Photos of people and vehicles in the United Statesaccording to materials reviewed by 404 Media and mistakenly disclosed by the company.

The findings raise questions about who exactly has access to footage collected by Flock surveillance cameras and where the people reviewing the footage might be located. Flock’s technology has become widespread in the United States, where its cameras are in thousands of communities and used by police every day to investigate things like car theft. The local police have too Conducted many searches for ICE in the system.

Companies using artificial intelligence or machine learning regularly turn to foreign workers to train their algorithms, often because the labor is cheaper than hiring locally. But the nature of Fluke’s work — creating a surveillance system that constantly monitors the movements of US residents — means the footage may be more sensitive than other AI training jobs.

Flock cameras continuously scan the license plate, color, make and model of all vehicles they pass by. Law enforcement can then search cameras across the country to see where else the car was driven. Authorities typically mine this data without a warrant, which prompted the American Civil Liberties Union and the Electronic Frontier Foundation to do so Recently sued the city Covered by nearly 500 Fluke cameras.

More broadly, Flock uses artificial intelligence, or machine learning, to automatically detect license plates and vehicles And peopleincluding the clothes they are wearing, from camera footage. Herd patent He also mentions cameras that detect “race.”

Several tipsters to 404 Media pointed to an online dashboard that showed various metrics associated with Flock’s AI training.

It included numbers around “annotations completed” and “annotation jobs remaining in queue,” where annotations represent notes that workers add to reviewed footage to help train the AI ​​algorithms. Tasks include classifying vehicle types, colors, and types, transcribing license plates, and “audio tasks.” Flock recently started advertising a feature That would reveal the “scream”. The panel showed that workers sometimes completed thousands and thousands of annotations over the course of two days.

The exposed board included a list of people assigned to annotate Fluke shots. Taking these names, 404 Media found that some of them were located in the Philippines, according to their LinkedIn and other online profiles.

Many of these people were hired through Upwork, according to the exposed materials. Upwork is a freelance platform where companies can hire designers and writers or pay for “artificial intelligence services,” according to Upwork’s website.

The guides also pointed to several publicly available Flock displays that explained in more detail how workers had to classify the footage. It’s not clear what specific camera footage Flock’s AI workers are reviewing. But screenshots included in the workers’ guides show numerous images of vehicles with US plates, including in New York, Michigan, Florida, New Jersey and California. Other images include road signs clearly showing that the footage was taken from within the United States, and one image contains an advertisement for a particular law firm in Atlanta.