One of my top priorities when I joined Arnold Ventures mid-2023 was to launch a standing, open Request for Proposals (RFP) for strong causal research related to public safety. Most research funding at AV had previously been invite-only, with occasional, targeted RFPs on specific topics. As a long-time academic, I’d become good at playing the networking game typically required by foundations (rules of thumb: find a way to meet program officers, and never submit a proposal without talking to them first), but I hated the opaque processes and relationship-based gatekeeping that pervades the philanthropic space. I was determined to create a more transparent process that made it easy for scholars to send AV their ideas. This reduces frictions, and is better for science. (I also wanted to streamline the RFP itself - no researcher wants to read 10-20 pages to figure out if their project is eligible!)
Fast forward through the hefty logistical lift of setting up a submission portal and external review process: this new RFP launch in the spring of 2024. The eligibility criteria were straightforward: we were looking for strong, causal studies measuring the impacts of US-based interventions related to crime, the criminal justice system, or people with criminal records. And all that was required of researchers was a 3-page letter of interest (LOI).
It’s been very fun to see the wide range of interesting projects that have been submitted since then. In line with my transparency goals, we collaborated with the AV Comms team to find a way to post grants made through this RFP online. You can find press releases with those lists on the AV website: 2024, 2025 Q1 and Q2, 2025 Q3, 2025 Q4, 2026 Q1.
(The releases are based on dates of finalized contracts — which could sometimes be much later than they were actually approved due to long bureaucratic delays on the university side. Related: I’ve always wanted to release a public list of average time from approval to finalized contracts, by university, to make information on such delays more visible to university leadership and prospective faculty hires. It’s truly insane how long some universities take to process and accept external funding; this slows down the work and drives both faculty and funders crazy. Some sunlight on all of this would helpful.)
Many funded projects are what I think of as “workhorse” research. The topics may sound less sexy (though of course I think all research is sexy!) but the projects will still produce useful evidence to guide a field that has, for too long, relied on anecdotes to make decisions. We need to provide more incentives to engage in such research; funding helps with this!
But some projects are especially exciting because they test a novel and scalable approach to long-standing problems. These interventions often involve new technology or medical innovations, or bring creative ideas from other fields into the public safety space. They can sometimes be moonshots: especially low-probability, but high-impact if they work.
Today and in future issues of this newsletter, I’ll share some of the funded projects I am most enthusiastic about, to give you a sense of the current research frontier and what I see as the most intriguing open questions in the public safety space.
First up: a new DNA analysis technique that has become a hot commodity in law enforcement circles.
PIs: Aaron Chalfin, Romaine Campbell, and Brandon Del Pozo.
AV funding: $4.1 million, in Q4 of 2025.
For this project, the PIs are working with law enforcement agencies in multiple states to measure the value-add of forensic genetic genealogy (FGG, pronounced “fig,” and sometimes referred to as investigative genetic genealogy, IGG) to identify suspects in criminal cases, relative to standard investigative practices. Using a randomized controlled trial at the case level, PIs will “horse race” FGG with what would have happened without it, in terms of the likelihood a suspect is identified, arrested, and convicted (after all, we want to get the right person, not just anyone), and the speed with which that happens. If our goal is to solve more crimes, faster — and it should be! — FGG appears to be a powerful new tool in our collective toolkit. But we need experiments like this to tell us just how powerful it is in practice.
FGG uses DNA analysis to identify perpetrators who are not (yet) in a law enforcement DNA database. (It is also used to identify unidentified human remains, which can provide useful leads in criminal cases and bring valuable closer to families.) It does this by using genetic information uploaded (with consent) by users of websites like ancestry.com, 23andMe, and GEDMatch, looking for full matches to crime scene evidence as well as partial matches that could point to a family member — sometimes a distant family member.
(If your DNA is already on one of those sites, you can opt-in to share it for FGG analysis. See this guide from the Sherry Black Foundation for more on how to become a “genetic witness” in this way. Othram also allows direct uploads to its website, dnasolves.com.)
Such analyses are typically done by private labs like Othram and Bode Technology, and they aren’t cheap — currently about $10,000 per case. So, police departments can’t afford to use FGG for everything. And, as with most new high-tech tools, there is also a potential privacy cost. Not everyone is comfortable using this new form of analysis, going beyond people added by law to government databases. We will need to have many thoughtful, candid conversations in the years ahead about when using FGG is worth it.
To inform those discussions, it will be crucial to have empirical evidence on the benefits of this tool. How much does it move the needle, in practice? Are we talking about a 5-10 percentage point increase in the likelihood of solving a currently-unsolved case, or a 70-80 percentage point increase? Most of us would be willing to sacrifice more for the latter. We’re probably also more willing to accept a reduction in privacy to solve a murder than to solve a car break-in. Does the value-add differ across crime types as well, or is it just the perceived cost?
With the advent of AI, we are barreling toward myriad important and ethically fraught discussions like this one. We will need to get good at working through the tradeoffs. Which tools have a meaningful public safety benefit? Which ones also bring privacy costs, and how large are they? And is there a way to implement each tool that mitigates those costs, to make them acceptable to communities?
This study is a big and exciting step forward in the conversation about FGG. I’m grateful to the law enforcement agencies that are willing to engage in this rigorous pilot, rather than simply assuming the tool will be worthwhile. What we learn in their jurisdictions will benefit the entire field. There is lots of work ahead, for all involved, but I’m eager to see the results (expected in 2030).
The Art of Gathering, by Priya Parker. I pride myself in being an excellent host and community builder, but want to become better at both. As I’ve written before, events can be super useful and productive, when done well.
Travel that Made Me — a new podcast by my friends at Indio Media (they also make the "Fighting Crime” podcast mentioned below). Each episode is a long-form interview with someone interesting and accomplished, reflecting on their life through the places they’ve been. A fun listen that’s perfect for summertime adventures.
In the latest episode of Probable Causation, I talk with UC Berkeley professor Manisha Shah about her fascinating research on intimate partner violence (IPV) in Tanzania. Most interventions aiming to reduce IPV focus on adults, and especially women. Manisha and her colleagues focus on adolescents, based on the hypothesis that it is easier to change someone’s relationship habits and choices when they’re young than when they’re older, and perhaps already tied to their partner via finances and children. These researchers also went beyond the standard focus on women to target men as well — recognizing that men are active participants in these relationships too, and everyone needs to bring their best selves to the table for relationships to be healthy. (This approach can often be unpopular among advocates, who want available resources to go to women and girls. We can’t keep thinking of this as a zero-sum game!) In this study, Manisha and her coauthors run randomized controlled trials testing interventions for young women and the young men they’re dating. The impacts were big and striking. Check out our conversation to learn more!
In the latest episode of Fighting Crime (produced by Indio Media for the Criminal Justice team at Arnold Ventures), host Cristina Quinn digs into the topic of DNA. (Yes, my team and I have been thinking a lot about DNA over the past few years!) Cristina talks with Ashley Spence from the DNA Justice Project, Tiffany Roy from the University of Leicester, David Rausch from the Tennessee Bureau of Investigation, Max Isaacs from The Policing Project, and yours truly. The episode discusses new technological innovations in forensic DNA analysis, like the availability of Rapid DNA machines that speed up analysis; how DNA profiles are generated for criminal investigations, and what information they include; how to think about privacy; and evidence showing that adding people to DNA databases has a dramatic deterrent effect on reoffending. Check out the full episode here:
I had a blast talking with Meredith Edwards about my new book, for The Curious Introvert podcast. You can listen/watch here:
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.