By Matt Cagle

We just got a pile of public records from the police department in Fresno, California. They show that in the last two years, Fresno police have been using several different brands of social media surveillance software – all without the public’s consent.

What’s worse, the ACLU has discovered that Fresno police have been using an especially offensive piece of software called MediaSonar, which encourages the police to identify “threats to public safety” by tracking #BlackLivesMatter-related hashtags.

Social media surveillance software comes in many forms, but it generally works by automatically scanning huge batches of publicly available posts on networks like Twitter, Instagram, and Facebook. This kind of surveillance can place people under suspicion simply for speaking their mind online.

Earlier this year, activists from the ACLU chapter in Fresno sounded the alarm and brought our attention to Fresno’s use of social media surveillance software called Beware. According to news reports, Beware’s mysterious algorithm was assigning “threat levels” to residents. Local advocates wanted to learn more. So we sent a Public Records Act request to find out how Fresno police were tracking social media, and what they did with the information they gathered.

In response, the police department sent us 88 pages of documents about its experiments with social media surveillance software. The documents raise a number of unanswered questions about social media surveillance at the local level. How many people are being surveilled? How long is their data being stored? How do these intrusive programs gather and interpret information?

Although we still don’t know exactly how the Beware algorithm works – and neither does the Fresno Police Department, apparently – we do know that it gathers information on a person’s publicly available social media activity, and assigns them a threat level of green, yellow, or red. Marketed as a source of insight for officers on the ground, this mysterious software can label people as threats based on inaccurate information. For instance, a Fresno city council member was recently incensed to learn that Beware lists his residence at threat level yellow.

Social media monitoring is part of a pattern of unchecked surveillance. It’s yet another surveillance tool being used without transparency or accountability. And it risks targeting communities that are already vulnerable to police misconduct – especially communities of color.

Which brings us to MediaSonar. Promotional materials from MediaSonar encourage the surveillance of hashtags like #blacklivesmatter, #dontshoot, and #imunarmed. In an email to Fresno Police, the company’s co-founder announced that these “keywords” could “help identify illegal activity and threats to public safety.” During the same time period that Fresno tested MediaSonar software, Black Lives Matter activists took to Fresno’s streets to call for reform.

The idea that people could be labeled as “threats to public safety” simply for objecting to police violence is chilling. But it’s not the first time – recently, the Department of Homeland Security and the Oregon Department of Justicewere caught secretly monitoring Black Lives Matter activists’ online activities.

Protestors shouldn’t have to wonder whether what they write online or post on social media will brand them as a threat in the eyes of law enforcement. The government shouldn’t be collecting a digital record of people’s lives. And police surveillance plans shouldn’t be rushed forward with the public left out of the loop.

As police departments across the country consider using intrusive new surveillance tools, here’s the thing to remember – we don’t have to stand for it. We have a right to determine how law enforcement acts in our communities.

And Fresno residents are right to demand reform, including the adoption of an ordinance that ensures the public has a meaningful chance to weigh in before decisions about this or any other surveillance technology are made.

Technology is changing the way we communicate and organize around powerful ideas. At the forefront of these innovations are movements like #BlackLivesMatter. While the way in which our society expresses itself is shifting, the principles of the First Amendment remain unchanged. Advances in technology are not an excuse for new forms of unaccountable surveillance.

Matt Cagle is technology and civil liberties attorney at the ACLU of Northern California.​

This post was first published in the Los Angeles Daily Journal on December 7, 2015. 

By Heidi Rummel and Peter Eliasberg

In the last 20 years, two new technologies have shed a revealing light on American criminal justice: DNA and video recordings. DNA testing has revealed more wrongful convictions than almost anyone had imagined. Widespread video technology has documented abuses of force by law enforcement that previously were discounted or denied. Both technologies promise to improve justice if embraced by prosecutors.
In a recent case in South Carolina, a bystander’s video recording revealed that a police officer executed Walter Scott in cold blood by shooting him multiple times in the back as he was running away and then lied to cover up what he had done. Scott was dead, and no action would have been taken against the officer without the video.

In some cases, however, the injustice does not end with the lying. Instead, officers’ false statements are forwarded to prosecutors who file criminal charges against the victim. Gabriel Carrillo’s case is a notable example. Carrillo was handcuffed and brutally beaten by sheriff’s deputies in the Los Angeles County Jail. The deputies and a sergeant then filed false incident reports claiming Carrillo had gotten one hand out of his cuffs and assaulted them.

Those reports prompted the district attorney to file a criminal case against Carrillo, who faced years in state prison if convicted. Carrillo luckily had an excellent lawyer who showed the district attorney photographs of Carrillo’s bruised wrists, clearly illustrating that he had been handcuffed during the beating. Faced with that gaping hole in the deputies’ story, prosecutors dismissed the case against Carrillo on the eve of trial.

The DA’s office later considered filing a case against the deputies involved, but concluded they had done nothing wrong. Fortunately, the U.S. Attorney looked at the same evidence and indicted and won convictions against the deputies and sergeant who had beaten Carrillo’s face to a bloody and discolored pulp and then lied about it. The jury foreman said after the verdict that the photographs made it an easy case.

Carrillo’s case is not an isolated incident. Investigators uncovered widespread corruption in the L.A. Police Department’s anti-gang unit in 1999. One officer in the Rampart Division testified in court that he and other officers lied in official reports to cover up the beating of a suspect in 1998. The officer was eventually charged with assault with a deadly weapon, perjury, filing a false police report and conspiracy.

Eventually, nearly 100 convictions were overturned after this officer revealed police misconduct and corruption in his unit.

These incidents demonstrate that we cannot blindly trust in the word of all law enforcement officers, especially in situations where the officers have strong incentives to lie. In use of force cases, officers may lie to cover up their own misdeeds, avoid employment discipline, or criminal prosecution. Moreover, if the DA’s office files charges against the actual victim, based on the lies of the officers, those charges frequently protect law enforcement agencies and local and state governments against damages. Winning a criminal case against a victim virtually eliminates any chance that victim will win any damages in a civil suit. In short, any institution that investigates itself should be subjected to the strictest scrutiny. Otherwise, great miscarriages of justice may result.

Preventing these injustices is not always easy, and we should consider a variety of possible reforms. For example, it may be necessary to appoint independent prosecutors to review cases of deadly force by law enforcement officers. There is an inherent reluctance and potential conflict of interest for DAs to investigate, much less file criminal charges, against the very officers they rely on so heavily to prove most of their criminal cases.

However, there are workable and practical reforms to prevent injustices such as the filing of criminal charges against inmates based on false statements by sheriff’s deputies. One result of a 2012 report by the Citizens’ Commission on Jail Violence was the installation of video cameras in the jails.

These cameras often provide the best evidence of what really happened before, during and immediately after a use of force incident. As the commission stated in its report, “cameras are an invaluable tool for objective accounting and insight” and “enhance the reliability of the investigation of [use of force] incidents.”

Given the recognition that cameras are a valuable tool for determining what happened in a use of force incident, the DA should not file a criminal case against an inmate, or a deputy, for an incident in the jails without first reviewing the video evidence, easily accessible from the jail video cameras.

Unfortunately, a Nov. 17 article in this newspaper, “Habeas petition claims sheriff’s department suppressed critical evidence,” described a case filed against two inmates suggesting that the DA does not uniformly take this approach. In that case, it appears the DA filed criminal charges against two inmates based on incident reports provided by sheriff’s deputies before actually reviewing videotapes of the incident.

Only after months of dogged effort by one inmate’s lawyer did the sheriff’s department and the prosecutor produce all the tapes of the incident. A review of those tapes persuaded the prosecutor to drop the criminal charges against one inmate because they contradicted the deputies’ statements.

While it is true that prosecutors can, and sometimes do, drop charges after obtaining additional evidence, there are still good reasons for them not to file charges in the first place until they review all available videotape evidence.

First, there can be serious collateral consequences for a person who has criminal charges filed against him or her, even if those charges are later dismissed. Second, once the prosecution has filed charges against the victim, it becomes nearly impossible for the victim to participate in an investigation of the officers. Finally, when it files charges, the state commits to the truth of those charges. There are many obstacles for prosecutors to turn around and acknowledge that they got it wrong because they did not have all the information. It is harder still for the prosecutor to do a 180 degree turn and file criminal charges against the same officer whose reports formed the basis for the original criminal charges against the victim.

Video cameras were put in the jails primarily to help deter excessive force. But the DA should use them to review the best evidence before rushing to judgment and filing charges against inmates based on what deputies say. This common sense approach is even more necessary in light of the incentives officers sometimes have not to tell the truth and the jail abuse scandal that has unfolded in L.A. over the last five years.

Until the DA’s office has a policy and practice of reviewing all reasonably available evidence — including videotapes of the incident in question — we cannot be reasonably sure that the rush to judgment has been replaced with a search for justice.

Heidi Rummel is former assistant United States attorney and is now clinical professor of law at USC Gould School of Law. Peter Eliasberg is legal director at the ACLU of Southern California.