LAPD tested social media surveillance

Voyager Labs did not provide evidence that predictions made by the software were accurate. (Quelle: IMAGO / ZUMA Press)

In 2019, the Los Angeles Police Department (LAPD) tested software to monitor people on social media, which can allegedly also predict crimes and detect extremist beliefs. That’s according to internal agency documents obtained by the nonprofit Brennan Center for Justice. The software, run by the company Voyager Labs, uses fake social media accounts, among other things, to examine people’s online activities and map their networks of contacts. It also records the activities of friendly user accounts, even if their owners are not under investigation.

The software is not only supposed to help solve crimes that have already been committed. The developer company also claims that the software can identify people who may potentially commit crimes in the future, based on political, religious and extremist beliefs.

The Brennan Center for Justice doubts the reliability of the software and warns that it discriminates against Muslims and other marginalized groups. It said that documents did not show exactly which parts of the software the LAPD ultimately used. However, police said they examined more than 500 user accounts and thousands of messages during the testing phase.

Muslims and marginalized groups discriminated against

The Brennan Center for Justice had obtained the police department documents with the help of a public records request. “The records […] offer a broader window into the typically secretive industry of social media monitoring,” the organization writes. The use of such products by police raises “serious concerns” about violating the First Amendment, which ensures freedom of speech, expression and the press.

A case study in the files showed that the analysis focused “strongly” on religious or Muslim topics, among other things. The Brennan Center rated the content in the example as “ordinary” and saw no indication of a particular propensity to violence. The algorithm had nevertheless provided the posts in Arabic with warning signs.

The organization doubts the vendor’s claim that the software can translate Arabic “instantly and completely” into 100 other languages. “Natural language processing tools have widely varying accuracy rates across languages, and Arabic has proven particularly challenging for automated tools,” it said. “And even literal translation of social media content often misses the important cultural context," the organization wrote.

Nevertheless, the program uses this information to create a risk assessment of the analysed users in the form of a color code. This is supposed to signal the “connection or affinity to Islamic fundamentalism or extremism”. A human review of the automated assessment does not take place.

Voyager Labs, he said, fundamentally equates legal beliefs with planning violent acts. “Even an accurate categorization of individuals with “ties” to “extreme” ideologies, whether underpinned by Islamic or any other beliefs, would provide no actionable information to law enforcement.,” the organization judges. Voyager does not provide proof that the system can actually reliably analyse a person’s ideology.

How it works

According to the British Guardian, Voyager Labs is just one of “dozens” of U.S. companies that offer social network searching to solve or predict crimes. The company itself sees its products as “search and analysis engines”.

The documents reveal that the software absorbs all publicly available information about a person or topic, such as posts and connections. Content published by contacts would also be captured, including status updates, images and geotags. The data would be analysed and indexed. Non-public sources would also be consulted.

The software creates an overall picture of a person’s social media presence. It visualizes connections to others and evaluates them. Customers can also be shown targeted indirect connections via interpersonal contacts between two people.

Collection of social media data

The pilot project with Voyager did expire in November 2019. However, the LAPD continued to use some of the technology. In 2021, further contract negotiations had been held about a collaboration. The outcome is unknown. A police spokesperson indicated to the Guardian that Voyager was not currently being used. However, in previous statements, police had indicated that social media was critical in investigations, public safety, and monitoring major events.

This is also evidenced by documents published by the Brennan Center for Justice in September. They show that LAPD officers are also supposed to ask for social media information and email addresses from questioned civilians, such as witnesses. It doesn’t matter whether the individuals are accused of a crime or not. Also found in the documents were the interview forms that LAPD officers were supposed to use for data collection. The practice, which continues today, was instituted back in 2015 under former LAPD Chief Charlie Beck. However, it was not public knowledge until the documents were released.

In recently released records, according to the Brennan Center, there is evidence that the LAPD has purchased or considered purchasing social networking surveillance software from at least ten companies over the past decade. The agency is considered a pioneer in introducing new technology to U.S. police, The Guardian writes. The LAPD, it says, has a large budget and is testing programs that will later be adopted by other agencies.

Voyager “uninvolved,” Facebook furious

Voyager absolves itself of any responsibility. “These are the responsibilities and decisions of our customers, which Voyager is not involved in at all,” company spokeswoman Lital Carter Rosenne told the Guardian. She added that they abide by the laws of the countries in which Voyager operates. It remained unclear whether reviews are taking place on how the software is being used. Rosenne stated, “We also trust that those we do business with are law-abiding public and private organizations.”

In response to the revelations, Facebook parent company Meta called on the LAPD to stop using “dummy” accounts and to stop collecting user data. The company sent an open letter to LAPD chief Michel Moore last Thursday, stating, “Our policies prohibit developers from using data obtained on our platforms for surveillance, including processing platform data about individuals, groups, or events for law enforcement or national security purposes.” The police department should stop all activities on Facebook that include using fake accounts, impersonating people and collecting data for surveillance purposes, he said. (hcz)