Solving Crime the Crowdsourced Way

Could the wisdom of crowds help to solve crimes?

Crime-solving in America

Today, in America, 1 out of every 3 murder culprits are not identified [1]. Despite advances in DNA testing and other technological improvements, that national statistic has not improved in over a decade. Some attribute it to a growing distrust between the police and public, which makes witnesses reluctant to identify suspects. Others argue that police agencies have quite successfully focused on preventing future crimes rather than dedicate resources to solving existing ones. And those resources can be costly. By one estimate, criminal investigations cost America 2% of our GDP [2]. However you look at it, solving crimes is labor intensive, costly, and we’re not exactly getting better despite technological advances. This seems ripe for a crowdsourced disruption…

So no crowd-based crime-solving models exist?

Crowdsourced crime-solving is not exactly a new idea. Intelligence groups have always used tip hotlines in order to get crowdsourced help about crimes, possible suspects, or evidence. For example, in 2007 the Boston Police Department launched a Text-A-Tip hotline. In April 2013, the month of the Boston bombing, the hotline received 333 texts [3]. And those 333 texted tips paled in comparison to the activity happening on Reddit, a crowdsourced online content aggregator. In the days after the Boston Bombing over 870 subscribers and 1,600 visitors contributed to, voted on, and analyzed information published on the Findbostonbombers subreddit [4]. They created mass photo dumps, analyzed photos and videos, conducted amateur forensics, and identified suspects. The Reddit content was also laced with racial and religious bias, incorrect information paraded as facts, and, tragically, horrendous witch-hunts of innocent bystanders [4]. The real culprits were never successfully identified by the subreddit. But just one month earlier, in March 2013, crowdsourced commenters on Gawker, another online blog, had successfully identified and help capture an assault culprit in NYC [5].

So there could be an opportunity here?

The enthusiasm of Reddit users, community based crime solving groups, and amateur crime solving in general show that there is a strong interest from the general population to participate in crime-solving – even when they’re not getting paid to do it. I’d like to propose a crowd-based business model based on a software platform connects users with police agencies looking for photographic or video evidence. This platform would create value through aggregating, sorting, and stitching such content together. It would organize the stitched content together by location, angle, time of day, perspective,  And it would capture value through providing evidence to authorized crime solving groups and private investigators for a fee. As an additional incentive to upload content, authorized crime solving groups can offer a bounty payout to users for evidence that is successfully used. As an additional revenue stream, crime solving groups can upload their own proprietary content and attempt to stitch it with user generated content. This service would augment traditional crime solving methods rather than replace it.

How is this different than the Findbostonbombers subreddit?

The purpose of such subreddits have historically been, as the name suggests, to find culprits and solve crimes. This platform seeks to simply gather information from content captured by private citizens via phones, web-enabled security cameras, drone footage, ambient listening apps etc. It does not allow users to view each others content, discuss potential evidence, or come to community based conclusions. Preventing such discussions is key to preventing the bias and witch-hunting that occurred in the Boston Bombing case. The platform would only accept evidence for which it can be reasonably confident the content was not altered. And it would maintain the anonymity of users submitting content, unless the user would like to be contacted.

Is all of this technologically feasible?

Recent advances in data processing, computer vision, and image rendering have made this kind of undertaking more feasible and cost effective. Other technologies that speak to the feasibility include:

Fourandsix – a software platform that identifies if a photo has been altered based on metadata

Empathic Media – a virtual reality company that uses traditional reported news and content to recreate crime scenes

Megamovie – a initiative by UC Berkeley and Google to crowdsource photos and videos of an upcoming solar eclipse to help scientists learn about the sun

Is any of this evidence legally admissible? 

Sometimes. Today, photo and video evidence is legally admissible so long as it is both relevant and authenticated [6]. In this business model, relevancy would be on the onus of the crime solving group. Authentication is currently achieved in a variety of ways, even by a witness who observed the content being depicted in the photo and video evidence.

Disclaimer: I am by no means a legal expert, professional crime solver, or expert in this subject matter. I am simply a curious Reddit reader interested in thinking how such technologies could be harnessed more effectively and maturely.




TripAdvisor’s FlyScore: Not yet harnessing the power of crowds


Waze: “Help[ing] people create local driving communities that work together to improve the quality of everyone’s daily driving”

Student comments on Solving Crime the Crowdsourced Way

  1. SUPER interesting! Overeager amateur sleuths and viral dissemination of inaccurate information and figure-pointing are definitely big concerns, as seen in the Boston bombing case. I’m curious how you think law enforcement officials will perceive such crowdsourced efforts. My understanding is that many of these “call in your tips” hotlines result in false information and wasted resources trying to chase down phantom tips. But perhaps 1% accuracy in crowd-generated tips is enough to make it worth it. In any case, I wonder if law enforcement officials will view these efforts as being an encroachment on their expertise and a waste of their time. I also worry that in this age of viral media content, mistaken (albeit, well-intentioned) finger-pointing can damage and ruin a person’s reputation within minutes. There may be a lot of second-order effects from these crowdsourced efforts, even if they lead to a “solved” case.

  2. Great post! I really like the intentional avoidance of creating an online community. It solves for the groupthink and witch-hunt problem.
    I think also – the law enforcement side of the platform doesn’t even need to disclose what crime or culprit they’re looking for. They can just post a date, time and location (e.g. March 20th 7am-10am in central park circa 72nd street), and ask users to submit any and all footage. It becomes more of a data collecting / stitching service, and encroaches less on the detectives’ line of work.

  3. Hi Bansi. Fantastic post! We definitely need some innovation in crime solving to ensure that justice is served as fast as possible. I have a question around incentivizing the group for crowdsourcing. One reason why reddit or other crowdsourcing platforms are very popular is the community and social aspect of it. People tend to contribute and talk to people about their insights or build on the insights of other. Your suggestion for crowdsourcing in crime solving does not have a community angle attached to it. Do you think that would impact the adoption of the platform?

  4. Hi Bansi! Fascinating idea. Without a community angle to this platform, how would you go about acquiring a vibrant and active community? Also, I’m guessing that crimes don’t necessarily occur in a high preponderance of places where amateur crime-sleuths are found – how limiting (if it all) would the evidence be if it was only submitted by those passionate about crime-solving?

  5. Great post – cool idea. I think the greatest source of change coming from such a platform would be the anonymity awarded to those submitting information. Many people have information about crimes, but are too scared to speak up due to threats from the culprit, or even a personal connection to the culprit (like they are a family member). This platform would allow for anonymous tips and a degree of separation from the police which might make coming forward easier for witnesses.

Leave a comment