This article appears in the November/December 2021 issue of The American Prospect magazine. Subscribe here.
Brad Carson wanted to do one big thing before leaving the Pentagon: make sure that the Department of Defense accurately counted the number of sexual assaults in the military.
It was 2015, toward the end of the Obama administration, and he had become acting undersecretary of defense for personnel and readiness, the department’s top human resources role. The position is considered one of the most boring in the Pentagon. But Carson, then 48 years old and a former Oklahoma congressman, was excited to take it on. He had been a progressive leader as undersecretary of the Army.
Officially, 6,083 service members reported sexual assaults that year, but Carson knew the real number was much higher. He believed that bureaucratic hurdles prevented victims from reporting. And saddled with a buggy database system, neither commanders nor Congress had a complete view of trends in offenses and victims in real time. Carson was convinced that the inaccurate sexual assault data was leading to bad policy.
And so, in the spring of 2016, Carson asked a group of software experts in the Defense Digital Service (DDS) to create a new, more functional database. “It was a terrific idea. It remains a terrific idea,” Carson said.
Four Defense Digital Service staffers set up meetings with military lawyers, victim advocates, Pentagon commanders, and key players in Congress and the Obama administration. A month later, they sent around an initial report. It was scathing. “First and foremost, there is extraordinary low trust in the Department’s sexual assault data and reporting among both internal and external audiences,” they wrote in the internal report obtained by the Prospect.
But when the team presented their ambitious approach, Secretary of Defense Ash Carter’s office cast the report aside. Carter’s chief of staff, Eric Rosenbach, took the report’s authors into the secretary’s conference room and raised his voice, according to those present. Chris Lynch, the director of DDS, reassigned his staff. Everyone who touched the report was told to move on.
“They completely buried it,” said one person with knowledge of the project.
To this day, many of the shortcomings the authors identified have not been fixed; untold thousands of assaults have gone unrecorded.
Carter and Rosenbach now direct an academic center at Harvard, and in speeches, articles, and a memoir, Carter has promoted an ethos of innovation. Lynch founded the military contractor Rebellion Defense; in September, venture capitalists invested $150 million in his two-year-old startup, despite it having few public contracts. Bolstered by a network of connected board members and advisers—including Carter himself—Rebellion has been valued at a startling $1 billion. No one ever publicly mentioned the report again.
The Biden administration has made sexual assault prevention in the military a big priority, but five years later the problem has gotten much worse. The story of the 2016 sexual assault data report shows that internally, the Pentagon had the same ideas that independent commissions have put forward—yet the military leadership has persistently fought back against change.
To this day, many of the shortcomings the authors identified have not been fixed; untold thousands of assaults have gone unrecorded. In September, the Department of Defense laid out a new process to address the problem, but some of the elements are not scheduled to be implemented until 2030. Eight senators from both parties excoriated Defense Secretary Lloyd Austin for this leisurely timeline. “The men and women who serve in our military,” they wrote, “cannot continue to operate another day, let alone another decade, under a chain of command that is unwilling or incapable of taking decisive action to address this epidemic.”
Congress may force the issue with regulatory changes, in particular by taking sexual assault prosecutorial decisions out of the military chain of command, in the defense budget in December. But the Pentagon, as the report elucidates, already knows how it can make progress. This year, the Government Accountability Office published recommendations on how to reform the way the Pentagon tracks sexual assaults. The main recommendation: Improve the tracking system. Exactly the change outlined in detail five years earlier by DDS.
TERESA BEASLEY UNDERSTOOD THE CONSEQUENCES of the Pentagon’s broken database intimately. She worked as a sexual assault response coordinator (SARC) at the Air Force Academy in Colorado Springs. For a decade, she spent hours each day with victims making their way through a convoluted system that she felt did not put them first.
Since she didn’t trust the military’s official database, she kept her own tally of assaults on the campus in an Excel spreadsheet. That paperwork helped her piece together a trend: Around Halloween each year, sexual assaults spiked. One year, six cadets reported being assaulted on a single day, she recalled.
But the system the military uses to track victims’ reporting, the Defense Sexual Assault Incident Database, was so “un-user friendly,” Beasley said, that no one at the academy had ever noticed the trend.
The database’s shoddy interface made Beasley’s work more difficult. “It’s got bugs, it’s slow, it isn’t intuitive,” she said. Some of the issues are technical: It obscures important data and doesn’t talk to other databases that report these incidents. The system doesn’t include victims who were sexually assaulted by family members or intimate partners (which comprise up to 18 percent of assaults), since those fall into the Family Advocacy Program. The cases of civilian victims are usually not entered.
If a user left a single box blank on any of the 14 different screens, it would cause the system to crash. Beasley interacted with victims all day long, but stepping away to take a call or to speak to someone coming into her office often meant losing an entire report. “Data is frequently lost even while being currently entered,” she said.
The current system is also prone to inaccuracy. Cases must be inputted within 24 hours of an assault (or 48 hours for service members who are abroad). In some cases, Beasley told me, the victim is still intoxicated or drugged, and she felt it was unethical to have them sign legal forms when inebriated. Many victims never get officially recorded since they didn’t want to sign their names on confidential filings at the end of the complex intake process.
Experts don’t rely on the official database since it only captures a fraction of victims.
There is also no way to log some important information. The “alcohol-involved” box, for example, doesn’t specify whether the offender or the victim was intoxicated. Other key factors don’t fit easily into the provided squares, like whether there were multiple victims or multiple offenders and the exact location of the assault.
Once in the database, a case cannot be followed through the military justice system. “Within the armed services, they cannot track from the beginning of an investigation of a crime to the end,” a congressional source told me. “We were told it’s too hard.” Punishments and convictions are not tracked, and if a case reaches a pretrial agreement, it is removed from the database.
Beasley had gotten used to the system’s defects and was determined to act upon her findings about Halloween. “I said, ‘Hey, we found a time of year where we need to up our training or to keep an eye on the cadets,’” she told me. But the academy leadership didn’t want to look into trends. “Officers think they’ll get fired if they have a lot of assaults,” she said.
But the only way to see those trends was in Beasley’s spreadsheets. “It’s like, if you were a police chief, and weren’t allowed to know what the crime history was in the district you’re going to be responsible for,” said Andrew Morral, a senior behavioral scientist at RAND who has led studies on risk factors for sexual assault in the military.
A senior official conceded that commanders have little sense of assault and harassment in their own command. “One of the areas that we have not had visibility is really within the units and at the installation level,” said Elizabeth Van Winkle of the Pentagon’s Office of Force Resiliency, at a Senate hearing.
Coordinators across the armed services shared Beasley’s frustration. Many, like her, had started tracking victims on their own spreadsheets and doing ad hoc analyses for their commanders, who would rather there be no reports at all. “There is always the question of, in my unit, are we reducing numbers of actual sexual assaults or are we reducing numbers of reports,” said one SARC.
The military’s own surveys show that as few as 23 percent of assaults are reported. Experts don’t rely on the official database since it only captures a fraction of victims. “You know, most sexual assaults are not reported, almost no sexual harassment is reported,” said Morral.
As time has passed, the rate of conviction has been going down, and only a meager number of commanders have been held accountable for alleged crimes that occurred under them. Ash Carter, Obama’s fourth and final secretary of defense, committed the department to ending retribution against those who report sexual assaults. Yet 64 percent of survivors who reported, according to a recent military survey, faced retaliation.
IN SPRING 2016, Brad Carson, whose wife is an expert on sexual assault prevention and a senior Pentagon attorney, was “deeply and very personally committed” to using his own high-level Pentagon position to address the crisis. But he realized military sexual assault was not a priority of his boss. The previous defense secretary, Chuck Hagel, convened weekly meetings with his deputies about sexual assault prevention, said Carson, but Carter didn’t hold such meetings.
Carson forged ahead with a focus on mapping the scope of the problem. He wanted to “create a new database to collect data on reports, investigations, and dispositions,” according to an internal planning document obtained by the Prospect. The new resource would be something that could easily extract info from PDFs, be quickly searchable, and allow leaders to make informed decisions about prevention.
“I can actually fix this, and I’m going to,” Carson remembered thinking.
Carson also had a bolder idea: “to establish an independent office that would collect data on sexual assault.” The digital project would go further than just fixing a database, and would revolutionize the military’s sexual assault reporting mechanisms with pretty basic technological fixes. “I can actually fix this, and I’m going to,” Carson remembered thinking.
The Defense Digital Service was a new unit within the Pentagon staffed by an enthusiastic group of data analysts and software engineers from the private sector. These were not military folks. They openly marveled that they had landed a job in the Department of Defense; one posted a photo of himself tinkering on a MacBook while sitting in a military helicopter.
The DDS director, Chris Lynch, was a Seattle techie with no military experience. He specialized in generating media buzz and networking, and he secured the blessing of Secretary Carter for DDS’s work. “He’s a salesman,” recalled a senior defense official who worked with Lynch.
Lynch carried with him to the Pentagon the bro culture of tech startups. On Lynch’s birthday, a junior colleague brought into the office a cake for him with a phallus illustrated on top in frosting. There was heavy drinking at happy hours. According to his former co-workers, Lynch was having an affair with a woman from Army Cyber Command who worked closely with DDS.
PHOTO BY CHRISTOPHER MICHEL/CC 2.0 ATTRIBUTION
Lynch took many of his DDS colleagues with him to start a military-tech company.
But the project had the positive glow of something new at the Pentagon. Carter brought Amazon founder Jeff Bezos to the DDS office “to see how tech is transforming government.” Through this and other events, Lynch fostered connections that would come in handy in a future move to the private sector.
Carson’s office reached out to Lynch, who assigned four young coders to research the data problem surrounding sexual assault. They were eager to get started. A former Google engineer wrote on the unit’s website, “I’m passionate about preventing sexual assault in the military, and I’m excited to get a chance to fix that problem.”
At the end of April 2016, Secretary Carter launched the project at a ceremony honoring six “Sexual Assault Response Coordinators of the Year.” There, he introduced the data initiative to a packed room in the Pentagon. DDS staffers sat in the fourth row.
“This project will help the department understand sexual assault data in a more meaningful way, also ultimately leading to more transparency between DOD advocates and others invested in this critical mission,” Carter said. “This will allow for more streamlined, timely, and accurate reporting.” Afterward, the techies ate slices of a sheet cake decorated with the sexual assault response coordinator logo, and talked about next steps.
But Carter would never mention the project again in public.
CARSON WAS NOT IN THE AUDIENCE. He had resigned suddenly a couple of weeks earlier. He had fallen out of favor with Congress and military leadership by proposing a series of aggressive personnel reforms that he called the Force of the Future.
Under Carson, morale in the office for personnel and readiness had been high. His staff didn’t mind sticking around till 9 p.m. or later. One aide even happily kept an air mattress under their desk for late nights. Many staffers felt the mood in the office quickly changed when the new acting undersecretary, Peter Levine, took over. Now, staff cleared out at 5 p.m.
Levine was a longtime Senate aide and a consummate bureaucrat. One appointee who worked with him described Levine as a “hatchet man” who was hired to undo Carson’s reforms. In a statement, Levine said he was brought in to get “over the finish line” reforms that Carson had initiated.
After joining in April, Levine told Lynch not to continue the initiative, according to one source. Levine disputed this. “It was my understanding that Mr. Lynch was working under the authority of the Secretary,” he said. “For this reason, I did not have the authority to tell Mr. Lynch not to pursue his project and I did not do so.”
Soon, DDS pursued a much narrower research project, or what startups call a “discovery sprint.” Secretary Carter signed a memo that gave DDS the authority to consult with a variety of officials, but they were told to not meet with experts at the White House and Department of Justice. They met with them anyway.
In May, DDS held briefings with a dozen policymakers, military leaders, and experts, including sexual assault response coordinators in each of the military services. In those meetings, almost everyone talked about the Defense Department Sexual Assault Prevention and Response Office (SAPRO) annual report to Congress. It’s typically packed with appendices of complex, almost indecipherable, graphs that run much longer than the report itself.
DDS staffers learned that administration officials and members of Congress were not satisfied with the report and the complex way it presented aggregate data. It was built off of bad information anyway, and the presentation made it even less useful.
But SAPRO wasn’t about to let an upstart office of non-military techies take over their core function. A turf war broke out over the discovery sprint. DDS was told to “tread lightly, ‘play nice’ with SAPRO,” said the knowledgeable person. The SAPRO director “was concerned that DDS was asserting the unilateral authority to establish a data system,” according to Levine.
DDS staff kept their heads down and continued their work. The authors focused on “technical solutions” and presenting “truthful data,” according to an internal DDS planning document that offered talking points for how to deal with their now-shrunken project. It ended with a call to action: “This is too important to screw up.”
IT WAS PAST 5 P.M., and the authors of the discovery sprint were called to meet with Rosenbach, Carter’s chief of staff. They gathered around a big table in the defense secretary’s office suite. Lynch was traveling in Afghanistan at the time, so the staff shuffled into the boardroom without their director. Rosenbach hadn’t slept for 48 hours, he told them, but had found a short window for a meeting.
Soon he was raising his voice at the members of DDS. He said they were “wildly out of scope.” They should have stuck with the technical problems of the slow, buggy data collection system, and nothing more. Rosenbach emphasized that they should never have written down their recommendations, according to two people with knowledge of the conversation. Once a report exists, he explained, it could get leaked. Carter popped his head into the meeting.
The DDS staff felt confused; they thought they were doing the right thing. It was intimidating for the coders who were doing brief stints in the Pentagon.
The authors wrote in plain English, in contrast to the jargony reports from advisory committees, which added power to the recommendations.
The report had proposed, “From the Secretary’s office down, the DoD must remove any perverse incentives to hide or affect sexual assault reporting.” That was a rebuke from the inside, making explicit years of neglect and even the active intervention of leading officials to prevent the appearance of a crisis. The authors wrote in plain English, in contrast to the jargony reports from advisory committees, which added power to the recommendations.
The Pentagon had long downplayed the mess, but here was a group of data experts noting that, on its own terms, the way the department tracked assault was insufficient, the quality of data was not high, and that “there is an extraordinary lack of meaningful dialogue between Congress and the DoD in the area of sexual assault prevention and response.”
After meeting with Rosenbach, the authors deleted those lines and wrote a more positive introduction, using a quote from Carter’s speech at the SARC of the Year event.
In the report, the authors offered recommendations that anticipated exactly the suggestions that expert panels would publish in the years to come. Some of the suggestions could help SARCs and military leaders better understand the trends within their ranks, like data tools to provide “localized analysis [that] will help identify destructive micro-climates within commands.” They advised “enhanced data collection” that would track cases throughout the military justice system, “from the very beginning to the very end,” which the current database doesn’t show.
DDS pointed out categories of sexual assault victims that SAPRO doesn’t report annually to Congress—maybe to keep numbers lower—like those under the jurisdiction of the family program, military criminal investigation, or local authorities. They also recommended including sexual harassment incidents, which were hardly being tracked. The authors urged the Pentagon to “embrace transparency and put itself on a path to public transparency.”
The appendix suggested new data categories to add to make collection more applicable and fixes to improve user experience. They explained that DDS was “ready to partner” with SAPRO to create a commander’s dashboard, where leaders could see real-time visualizations. “I think it is major,” said a Pentagon official who was familiar with the report. “Just imagine if you had a commander’s dashboard that DDS proposed.”
For Levine, data collection was not simply a technological problem, but a policy problem with potential legal repercussions. “In my view, DDS was in no way expert on where the lines were,” said Levine. He told me he had never seen the DDS report.
Lynch declined interview requests. SY Lee, a spokesman for the company he has since co-founded, wrote in an email, “Chris Lynch and the team at DDS regarded sexual assault reporting in the military as a top priority.” Rosenbach declined to officially comment, and Carter did not respond to the Prospect. None of the four authors of the report wanted to be quoted.
Lynch quietly reassigned some of the authors to other projects; others departed DDS entirely. When asked whether there were any next steps for the sexual assault report, Lynch said to a DDS staffer, “No,” in a way that intimated that the question shouldn’t be asked again.
Manuel Balce Ceneta/AP Photo
Secretary Austin has readily acknowledged the lack of progress on sexual assault prevention.
TERESA BEASLEY, the sexual assault response coordinator, left the Air Force in 2017 after 30 years, fed up with what she saw as an intentional effort to conceal sexual assault trends. The academy’s leadership didn’t believe her when she reported back to them an upward trend in the number of sexual assaults on campus. Beasley filed a whistleblower report and was forced out. According to the Office of the Inspector General, her bosses had indeed doctored the numbers to lower the count that would be reported to Congress.
“For people who think data doesn’t count, it does. And accurate data really matters,” she told me. “These aren’t just numbers. There are real stories behind each one.”
Little has changed in how the military tracks sexual assaults in the meantime. Though the Trump administration continued to charter independent reporting bodies and committees, his secretaries of defense did not prioritize the issue.
Experts have advised urgent fixes to data collection that reinforce DDS’s findings. An independent federal commission recommended in 2020 that the Pentagon “develop a single electronic database for the uniform collection, storage, and analysis of standardized military justice documents across the Military Services,” but there has been no movement on this front.
When a 20-year-old private, Vanessa Guillen, was found dead and dismembered at Fort Hood in April 2020, the secretary of the Army launched an investigation that described a “permissive environment for sexual assault and sexual harassment.” Watchdogs saw Guillen’s murder as part of a bigger trend of commanders not focusing on prevention.
Updating the military justice system is a major priority for those seeking reform. Joe Biden throughout his career had advocated for sexual assault victims and on the campaign trail said “yes, yes, yes” to moving the justice system away from the chain of command. Secretary Austin, in his Senate confirmation hearing, readily acknowledged the lack of progress on sexual assault prevention and months later supported removing prosecutorial power from commanders. The 2021 Independent Review Commission on Sexual Assault in the Military made this issue its signature policy change, but the timelines for implementation are exceedingly slow. (Biden supports the commission’s recommendations.)
“The collection and reporting of data should be baked into policy change,” said Lynn Rosenthal, who oversaw the commission. “There should be greater transparency, more accessibility of the data. That’s the only way we’ll know whether this policy change made a difference or not.”
Five years after the DDS report, this year’s defense funding bill may at last put questions of justice outside of the commander’s reach and into the hands of military judges and an impartial justice system. Sen. Kirsten Gillibrand (D-NY) has built a coalition of 66 senators for an amendment that would apply to all felonies, not just sex crimes, and it has the support of both Democrats and Republicans, including Minority Leader Mitch McConnell and Joni Ernst, herself a veteran and a sexual assault survivor. But Senate Armed Services Committee Chairman Jack Reed (D-RI), who has close ties to the defense establishment at the Pentagon, is a holdout. He may bring down an otherwise popular piece of legislation.
“The existence of this report makes it clear that the Defense Department knew about its sexual assault data collection problems years ago,” said Sen. Gillibrand.
Carson believes a data-driven report in 2016—a year before the #MeToo movement—could have accelerated meaningful reforms. “It might have actually been a story that put sexual assault in greater context,” he told me. “It wouldn’t be a negative story for the military. It would allow us to have better, more acute interventions.”
“The existence of this report makes it clear that the Defense Department knew about its sexual assault data collection problems years ago, but chose to bury them rather than address them,” said Sen. Gillibrand in statement. “This report serves as yet another reminder that the military will take every opportunity to drag its feet when it comes to changing its approach to sexual assault.”
The DDS report was further validated by a memo Secretary Austin issued in September, a road map that integrates many of the GAO recommendations and that responds to the Independent Review Commission. An entire section on data explains how the services will modernize: They will integrate domestic abuse and intimate partner cases into the database, add new categories to data collection, and deploy a “tool that would enable unit-level commanders to collect real-time climate data on sexual harassment and sexual assault,” which is what DDS recommended.
But there’s a catch: “Estimated implementation by 2028.” The timeline is now 12 years outside of what DDS proposed to have done in 2016.
WHEN LYNCH LEFT THE PENTAGON in April 2019, he took many of his DDS colleagues with him to start a military-tech company. He named it Rebellion Defense after the rebel forces in Star Wars. Rebellion is marketing AI products for military and intelligence.
Lynch’s revolving-door connections helped grease his path, through an interlocking set of influential people. The former secretary of defense, Ash Carter, advised the new company. (He has no financial interest in Rebellion, according to the spokesperson.) At the same time, Carter and Rosenbach took up jobs at the Harvard Kennedy School of Government. Among Carter’s roles at the university is overseeing a tech-policy fellowship program sponsored by ex-Google CEO and billionaire Eric Schmidt, who is also a key investor and board member of Rebellion. Lynch holds a prestigious fellowship in Carter’s Harvard program. (“Chris Lynch followed strict DOD guidelines as he transitioned back to the private sector,” said the Rebellion spokesman.)
The mass exit to Rebellion has left the Defense Digital Service understaffed and disorganized, Pentagon officials told me. Morale suffered under Lynch’s successor, Brett Goldstein. In the spring, 30 staff complaints were made about Goldstein’s leadership and the way he treated women in the workplace, which triggered an investigation. The inspector general could not substantiate claims of a “negative work environment,” and Goldstein left DDS in June 2021.
Lynch founded Rebellion Defense with the colleague with whom he was having a romantic relationship in the Pentagon. An unwelcoming bro culture endures under Lynch in the private sector, say well-placed sources. (The Rebellion spokesperson says the company has the “highest standards of inclusivity and transparency, with zero tolerance for unprofessional behavior.”) Meanwhile, a Rebellion colleague recently participated on an all-male panel titled “The Importance of Male Allies.”
Lynch loves to talk about making technology for the “warfighter” in interviews and events, but he has kept quiet about efforts to address deeper social problems that the warfighter faces, like sexual assault.
Sometime in 2018, sources explained, at a reunion with co-workers and friends from DDS and the U.S. Digital Service, Lynch held court at a bar in downtown Washington. Some of the coders were back in the private sector, and still in shock that they had been spending their days sprinting down the Pentagon’s halls or fixing tech problems for the White House. As usual, they all got a little too drunk.
On the street afterward, someone asked Lynch what had happened to the sexual assault report. They wanted to know why a report that had seemed to solve so many policy problems seemed to have vanished.
They remember Lynch saying, “That’s the only failed project DDS ever had.”
