The Problem with Crowdsourcing Crime Reporting

There has been some excitement about the idea of using technology to address the problems of the Mexican Drug War. As someone involved in technology, I find it inspiring that other techies are trying to do something to end the conflict. However, I also worry when I read ideas based on flawed assumptions. For example, the assumption that “good guys” just need a safe way to report the “bad guys” to the cops reduces the Mexican reality to a kid’s story, where lines are easily and neatly drawn.

So, here are a few reasons why building tools to enable citizens to report crime in Mexico is problematic and even dangerous.

  1. Anonymity does not depend only on encryption. Criminals do not need to rely on advanced crypto-techniques when information itself is enough to figure out who leaked it. Similar ideas are being discussed by researchers trying to figure out how to identiy future Wikileaks-like collaborators, something they call Fog Computing. The point is, the social dynamics around the Drug War in Mexico mean that people are exposed when they post something local. In an era of big data, it’s easy to piece things together, even if the source is encrypted. And, sadly, when terror is your business, getting it wrong doesn’t matter as much.
  2. Criminal organizations, law enforcement, and even citizens are not independent entities. Organized crime has co-opted individuals, from the highest levels of government down to average citizens working with them on the side– often referred to as “halcones.”
  3. Apprehensions do not lead to convictions. According to some data, “78% of crimes go unreported in Mexico, and less than 1% actually result in convictions.” Mexico is among those countries with the highest indices of impunity, even with high-profile cases such as the murder of journalists.  All this is partly because of high levels of corruption.
  4. Criminal organizations have already discovered how to manipulate law enforcement against their opponents–there is even a term for it: “calentar la plaza“– the sudden increase of extreme violence in locations controlled by the opposite group, with the sole purpose of catching the attention of the military, which eventually takes over, and weakens the enemy.

The failure of crowdsourcing became evident only a few weeks ago with a presidential election apparently plagued with irregularities. Citizens actively crowdsourced reports of electoral fraud and subsequently uploaded the evidence to YouTube, Twitter, and Facebook. Regardless of whether those incidents would affect the final result of the election, the institutions in charge seem to have largely ignored the reports. One can only imagine what would happen with the report of highly profitable crimes like drug trafficking.

Crowdsourcing is not entirely flawed in the Mexican context, though. We have seen people in various Mexican cities organize organically to alert one another of violent events, in real time. But these urban crisis management networks do not need institutions to function. However, law enforcement does, unless one is willing to accept lynching and other types of crowd-based law enforcement.

In sum, as Damien Cave mentioned, what Mexico needs is institutions, and the people willing to change the culture of impunity. Technologies that support this kind of change would be more effective than those imagined with a “first world” mindset.

Thanks to danah boyd for helping me think through some of these ideas.

Shouting Fire in a Crowded Hashtag

Narco Censorship

The press is one of the many casualties of Mexico’s ongoing violence, in particular, the local media. Newspapers and TV stations are caught in a battle between censorship, control and threats from the drug cartels, and the local governments. In some cities, people often witness shootings, grenade attacks and other violent events, but when they try to find out what happened, their local news has nothing to offer. Some newspapers have officially announced a policy of self-censorship when it comes to reporting drug war-related news.

The result for a lot of Mexicans is that local media is no longer a source of news. Some citizens claim that their local news sources are paid off by the local government in an effort to minimize the violence; others argue that it is the cartels who have bribed them; while others, especially the journalists, say they are being threaten to stay quiet. What is certain is that journalists are being murdered and their murders often go unpunished.

Hashtags Save Lives

Knowing if there is a shooting going on in a certain part a city, is not just about satisfying one’s own curiosity, but about one’s safety. Many of the recent violent episodes in Mexico last long enough that knowing about them can be a life-saving piece of information. Since main stream media no longer fulfill its role of informing citizens about these events, people have turned to social media.

Twitter in particular, with its unidirectional follower model and its hashtags, has become one of the main sources of citizen-driven news in Mexico. People often “report”, “confirm” and re-tweet information about violent events using hashtags. In several cities, hashtags have emerged as shared news resources. One of the first cities where these hashtags were used was Reynosa, with #reynosafollow, followed by Monterrey with by #mtyfollow and, more recently, #verfollow for the coastal city of Veracruz.

A word count analysis of more than a quarter of a million tweets using the hashtag #mtyfollow over the course of nine months (11/2010 to 8/2011) shows how hashtags are used as a common resource. People hook into the hashtag to “report” (“REPORTAN”, in Spanish), issue warnings (“precaución”, “cuidado”) and request confirmation (“confirmar”) about shootings (“balacera”, “detonaciones” “balazos”) in certain areas of the city (“zona”, “Cumbres”, “Av”, “Sada”). You can also see the popularity of some user handles in the messages. Together, people such as @trackmty, @AnaRent, and @cicmty, have more than 85,000 followers and 65,000 tweets. These people have become reliable information news sources.

Most common words in 252,431 tweets using the hashtag #mtyfollow

Twitter Terrorists

Last Thursday at 11:56 AM, @gilius_22 tweeted a message using the #verfollow hashtag. He claimed that five kids were kidnapped at a school:

#verfollow I confirm that in the school ‘Jorge Arroyo’ in the Carranza neighborhood 5 kids were kidnapped, armed group, panic in the zone

The message was re-tweeted by twelve people, one of them was @VerFollow, a popular account with more than 5,000 followers that was created to report on the violence in the city. Immediately after these tweets, the rumor started spreading like wild fire. There were reports saying that one the of drug cartels was threatening to kill a child for each cartel member killed. People spread the news via Facebook, emails, and text messages. @gilius_22 reported that the cellphone network went down. Additionally, several other twitter users reported other incidents related to schools and to helicopters supposedly flying at low altitude.

By 12:00 pm (only four minutes later) the governor tweeted a message dismissing the rumor. However, by then it was either too late or the governor was not considered a reliable news source (probably a bit of both). Many parents rushed to to pick-up their children from school, causing massive traffic, chaos and panic across the city. Many parents did not take their kids to school the next day and businesses reported a 70% productivity loss due to the incident.

Mentions of the hashtag #verfollow in the month of August. Note the spike in Aug 25, the day of the rumors. Source: Topsy.

By 12:05 pm the governor tweeted his support for freedom of expression but urging people to make sure information is reputable before acting on it. Three hours later he posted that the government would go after those who spread the rumor on the basis of “terrorism”:

We have identified today’s misinformation sources, I want inform that this will have legal consequences according to Article 311 (terrorism)

Wikileaks-inspired logo of the anti-censorship movement in Veracruz.

The same day, the government website issued a statement listing sixteen twitter accounts involved in the rumor and threatening to take legal action against them. The statement also mentioned the name of the person associated with the account @gilius_22. By Saturday, @gilius_22 and @maruchibravo were arrested on charges of terrorism. Today, the total number of arrests has increased to three. Some of them have claimed to have been tortured by the police and forced to sign confessions. At the same time, many Twitter users across the country have rallied in opposition to the arrests. Many have mocked the government  by calling themselves “twitteroristas”. There is even an Anonymous video  “denouncing the government’s reaction against social media and the “lack courage” of the local media to report what is happening in the city.

Social Media Fail?

It is unclear what the motives and roles were of those sixteen people charged with spreading the rumor. Did they shout fire because they thought they saw flames or did they completely invent it? What led to the fast viral spread of this rumor?

The rumor would not have spread as easily if there was not already a widespread sentiment of vulnerability. It is unclear what did happen that day. There are several reports of military mobilizations around the same time of the tweets. If that was true, it probably added legitimacy to the rumors. Shouting fire in a theater carries a lot more weight that shouting fire in a pool.

The rumors spread faster because of a weak information “immune system.” Main stream media and the government are no longer considered reliable information sources in some of these cities. Social media has taken the role of the main stream media and that comes with its own challenges. Social media (i.e. Twitter) has fluid reputation mechanisms, which is positive because it helps protect people’s pseudonymity in light of the real danger faced by journalists. On the other hand, these fluid reputation mechanisms are problematic for assessing the reliability of information.

Many citizens do not trust the government. For example, the official Twitter account created by the local government to report violent events had six times less followers than some of the citizen journalists on Twitter. Many people claim that the government often downplays or completely denies the existence of any kind of violence under the motto “no pasa nada” (“nothing happens”). The governor himself has explicitly denied saying such thing:

I have never said that in Veracruz ‘nothing happens,’ we are fighting crime with all of our power so we can live in freedom, that is what is happening.

The circumstances were fertile ground for spreading misinformation. However, prosecuting Twitter users raises some questions. Yes, their actions caused panic, but does it actually amount to terrorism? Also, it is likely that these arrests will have a chilling effect on social media in Veracruz and maybe other cities, destroying citizen’s last resort for news. Another possible outcome is that social media might be pushed underground, making it even harder to develop reputation-building mechanisms.


If you found this interesting, you can follow me on Twitter or identi.ca.
Thanks to Nick Diakopolous for his feedback on this post.

UPDATE: Related interview on CNN and RWW article.

How much is a life worth in pixels?

Analysis of yesterday’s news coverage of the Mexican massacre

Mexican Tweets

More than fifty people were murdered yesterday in what is now the most violent episode in the ongoing Mexican Drug War. Most of the victims were women, some were pregnant. After learning about the horrific massacre in Monterrey, I spent several hours reading the reports coming from México via social and mainstream media. I exchanged messages with friends and family who live there (I went to college in Monterrey and my parents live no too far from there). The Twitter trending topics in México showed anger, desperation and hopelessness. One of the hashtags people often use to report violence in the city, #mtyfollow, was full of messages of repudiation and of people trying to help others find their loved ones. Some of the most retweeted messages were those with the names of the possible victims, as you can see in this chart.

 Twitter activity on a popular keyword right after the massacre
Mexican Twitter users helping find missing people after the massacre

American Silence

The massacre  happened only  140 miles south of Texas in one of the largest metropolitan areas in North America. Yet, as Nancy Baym put it,  the American twittersphere was mum. Why? In part, I think, because most of the news websites in the US were ignoring the event.

One could understand the lack of coverage in the first few hours. The news coming out of México were talking about “only” four deaths, so it is possible the events might not have caught the attention of the American news websites at first. However, ten hours after the attack the official number was already above fifty victims, with some reports as high as 61, yet sites like CNN.com gave little attention to the story. The link to the article of the massacre was buried among articles such as one about actress Rose McGowan’s childhood.

I know CNN is not known for its high-quality news coverage so I decided to check out one of America’s most trusted news outlets:  the New York Times.  I was disappointed, again.  I had to scroll all the way down to the “More News” section to find a 10 pixel-font link to the article titled “Arson Kills 40 in a Casino in Mexico.”

Pixels per Victim

Frustrated by this, I decided to get a more objective assessment of the coverage by counting the number of pixels different news websites were assigning to the story of the massacre. I know web designers put a lot of work into every single pixel on the screen, especially of high-traffic websites. Visitor’s attention is scarce and every pixel counts. So I took screenshots of  the front pages of some of the major news websites and calculated the amount of screen real state assigned to the story of the massacre. For example, the the New York Times, gave the story 291×11 pixels, a mere 0.27% of the screen real state (in a window size of 1439 x812 pixels). CNN gave it even less at 191×10 pixels, representing 0.16% of the screen. But what about other websites? Did any other websites in the English-world gave it more space? Yes. Read on.

I decided to look into non-American websites. If my calculations are correct, it turns out that Al-Jazeera and The Guardian alone gave more pixels to the story than CNN, the Washington Post, FOX News, the Wall Street Journal, the New York Times, MSNBC and the Houston Chronicle combined. Americans might be better off getting news about  their southern neighbor from a British or a Qatari website than from many of the US ones. The two exceptions were the LA Times and the Huffington Post. They both gave more pixels to the story than any other news source I analyzed. CNN was at the bottom of the list though. Click here for a slideshow of the websites I analyzed.

To summarize my results, I generated a ranking of the number of pixels per victim each news website devoted to the massacre. Yes, this issue is much more nuanced than pixels per victim, and I am not a journalism expert but I hope it can help start a discussion (or continue an existing one). If my calculations are correct, CNN devoted 38 pixels per victim, 76 times less than the LA Times which gave 2,920 pixels per victim.

Closing Thoughts

The Mexican Drug War is a complex geopolitical conflict closely linked to the United States’ financial stability  and national security. If American news websites do not give enough attention to the massacre of 50 people, what can we expect of less dramatic stories with perhaps more structural and long-term implications? I list here some of the recent related stories that I wish had gotten much more attention and that I hope you get to read to understand the complexity of the problem:

  1. The Guardian’s article on “How a big US bank laundered billions from Mexico’s murderous drug gangs.”
  2. The LA Times’ article on a senate report on how the “U.S. can’t justify its drug war spending” (there are many more articles about this).
  3. The NY Times story on how US-officials “allowed nearly 1,000 guns to flow illegally into Mexico” (also check this campaign to stop gun smuggling).
  4. Chomsky’s excellent synthesis of the whole Drug War problem  with a historical perspective that only Chomsky can give.


If you liked this, follow me on Twitter or identi.ca.