Ever since the U.S. Supreme Court ruled in 1986 that sexual harassment violates Title VII of the Civil Rights Act, vindicating Mechelle Vinson, a bank teller fired after a higher-up subjected her to years of unwanted sexual aggression, many employers have behaved as if the most important aspect of workplace culture is reducing or eliminating exposure to liability.

That paradigm helped some victims of sexual harassment to recover damages, and it prompted changes that presumably spared others from being victimized.

But its shortcomings and inadequacies have been laid bare by the allegations against Roger Ailes, Harvey Weinstein, and others whose alleged misbehavior was exposed or anonymously described in the #MeToo campaign. “I’ve received somewhere between five and 20 emails every day from women wanting to tell me their experiences: of being groped or leered at or rubbed up against in their workplaces,” Rebecca Traister, one of the most eloquent chroniclers of this outpouring, wrote at The Cut. “They tell me about all kinds of men—actors and publishers; judges and philanthropists; store managers and social-justice advocates; my own colleagues, past and present—who’ve hurt them or someone they know. It happened yesterday or two years ago or 20. Few can speak on the record, but they all want to recount how the events changed their lives, shaped their careers; some wish to confess their guilt for not reporting the behavior and thus endangering those who came after them.”

She went on to describe one inadequacy of the movement and the journalists covering it: “There are also women who do want to go on the record, women who’ve summoned armies of brave colleagues ready to finally out their repellent bosses,” she noted. “To many of them I must say that their guy isn’t well known enough, that the stories are now so plentiful that offenders must meet a certain bar of notoriety, or power, or villainy, before they’re considered newsworthy.”

If #MeToo at its height cannot give a hearing to most victims of serious abuse, or trigger accountability or change in any industry sufficiently removed from the creative class, what happens when it fades from social media, headlines, and public consciousness? Absent a structural change, most employers and human-resources departments of the future are likely to remain inadequate to the problem—a judgment seemingly shared by the anonymous creators of the “Shitty Media Men” list, who sought to formalize the “whisper networks” that have long existed in different industries to warn against abusers outside of  official channels.

Their approach was untenable, as observers quickly realized: The ability to publicly and anonymously accuse others of serious misconduct invites abuses and unjust outcomes, especially if it encompasses people unknown to one another personally—the scale necessary to effect significant reductions in sexual harassment.

Yet a different kind of third-party information sharing may still be possible.

A scholarly article published in 2012 by Ian Ayres and Cait Unkovic defined the challenge: Many are reluctant to be the first person to accuse someone of sexual harassment, in part because the accused “routinely responds by trying to impeach the credibility of the accuser.” Yet first accusations often lead to more accusers coming forward. That’s a dynamic that tends to protect recidivist harassers.

What if a system of “information escrow” existed instead?

We propose the use of an allegation escrow to allow victims to transmit claims information to a trusted intermediary, a centralized escrow agent, who forwards the information to proper authorities if (and only if) certain prespecified conditions are met. The escrow agent would keep harassment allegations confidential, unutilized, and unforwarded until the agent has received a prespecified number of complementary harassment allegations concerning the same accused harasser. For example, if the escrow agreement specified the accumulation of two additional allegations as a triggering event, then the agent would wait until the escrow had received three separate allegations concerning a particular alleged harasser before forwarding the information to specified authorities and initiating a complaint.

A variation on that idea is already being used by the nonprofit organization Callisto, a third-party reporting system for victims of sexual assault on college campuses.

Jessica Ladd, the company’s founder, built Callisto after extensive consultations with students who’d been frustrated by the process of reporting that they were sexually assaulted. It is already being used on campuses including Stanford, the University of Oregon, USC, and Pomona College. It offers three options to students: They can save time-stamped written accounts of a sexual assault; report the allegations electronically to campus authorities or police; or report the assault only if another victim names the same perpetrator.

The intent is to increase reporting rates and identify repeat offenders. Its website states, “15 percent of sexual assault survivors who opted into our matching system have been assaulted by the same perpetrator as another survivor in the system.”

Sooner or later, a similar approach will almost certainly be tested on workplace sexual harassment. Dozens of variations are possible. For example, imagine that a worker in most any industry could choose to report unwanted sexual behavior to a third-party sexual-harassment clearinghouse.

It might permit accusers a number of options:

  • Submit a sealed, sworn affidavit to document an instance of harassing behavior in the moment, without deciding to take any further action at that moment.
  • Submit a sworn complaint about workplace harassment that is sealed until some number of other people file similar complaints about the same company or individual, triggering notifications to all the victims.
  • Submit a ticket that alerts an HR department that an anonymous employee, verified to be working at their company, is concerned by an aspect of workplace culture.
  • Submit a ticket that generates a private alert to an individual notifying him or her that an unnamed co-worker asks that they voluntarily change a behavior.

Carefully designed and administered, a system like that might have stopped alleged serial abusers like Roger Ailes and Harvey Weinstein years ago by giving their victims access to a transformative insight: that they were far from alone. It might offer victims who dread the idea of going to HR but who also fear that others might be harmed if they stay silent an empowering option. Its mere existence would surely be a deterrent to some serial workplace harassers. And it might be a constructive way for well-intentioned people who cluelessly make a colleague (or several) uncomfortable to grasp how their actions affect others, without the need for an awkward confrontation or a formal intervention by management.

There are probably unintended consequences or vulnerabilities to abuse that I’ve failed to anticipate; I may be overestimating benefits or underestimating costs; and there are likely tweaks that would improve even the best of my suggestions.

When I ran them by Lara Stemple, the director of the Health and Human Rights Law Project at UCLA, she liked that my design anticipated low-level abuses and offered a range of responses. “I think there’s a huge issue right now concerning the breadth of definitions,” she said. “Smaller infractions are at risk of being treated like large ones, creating huge amounts of uncertainty that threaten to undermine the legitimacy of new efforts.”

But she also had reservations. “The only worry I have about anonymity would be along the lines of the toxic workplace culture that Amazon reportedly created with its anonymous reporting about co-worker performance,” she added.

Readers will surely think of other shortcomings, either with my specific thoughts, or with “information escrows” more generally. But there’s little question that there’s a pressing need for devising structural reforms to address workplace harassment.

#MeToo has been cathartic for millions who’ve been mistreated; it has prompted overdue, guilty introspection among some who’ve behaved badly; and it has been educational, constructive, and empathy-building for many besides. But as a mechanism for accountability, it relies so heavily on public shaming and stigma that its time horizon is limited, as are the number of victims who can be heard, especially beyond the privileged classes. And it remains extremely vulnerable to excesses and subversion, making abuses and cultural backlash likely.

It is made more vulnerable in part by the standard some of its adherents are suggesting: not Hillary Clinton’s insistence that “every survivor of sexual assault deserves to be heard, believed, and supported,” which is of course true, but the more extreme claim that because false accusations of sexual misconduct are atypical—and they are—everyone who makes an accusation must be believed, not only by their supervisor or the police, but by literally everyone.

That standard is untenable.

It is untenable because the prejudices and impulses that have led to pogroms and lynchings are not gone from humanity, and still put marginalized groups at special risk; because if adopted, the Duke lacrosse team would now be wrongly imprisoned; because when adopted, it caused Rolling Stone to be sued for libel, with an administrator at the University of Virginia among its victims; because if adopted going forward, the most vile political operatives would generate false accusations against senators in states where the governor belongs to a different party; and because even absent excesses of that sort, adherents sympathetic to most accusers, like Lena Dunham, who once tweeted, “Things women do lie about: what they ate for lunch. Things women don’t lie about: rape,” will sometimes be inclined to defend a particular accused person.

The ways Western culture has often treated accusers of sexual misconduct, whether women in the workplace or boys in the Catholic church, is abhorrent and unacceptable. But the flaw in those historical norms wasn’t a reluctance to believe all accusations—rather, it was the tendency to disbelieve almost all of them.

The remedy is not a new norm that will be undermined or discredited if even a single alleged victim is shown to have lied. Such a norm is setting accusers up for certain failure. Perfection is far too much to ask of a population of hundreds of millions.

But there is a new norm that could survive scrutiny and remedy the historic injustice: “Don’t reflexively disbelieve allegations of sexual misconduct or abuse.” Or, “Don’t reflexively disbelieve women accusing men. Every person deserves to have their claims of abuse heard and investigated with due diligence.” Using that norm, The Washington Post has published numerous accurate accusations after verifying their credibility, but did not fall for a trap sprung by a right-wing provocateur who recruited a woman to make a false accusation.

Of course, even that tenable norm will never be embraced with unanimity, as many Donald Trump and Roy Moore supporters are demonstrating every day. But who needs unanimity? If a majority of Americans simply decline to reflexively dismiss sexual-harassment allegations, even as accusers are empowered with an effective information-escrow system, alerting victims to one another’s existence as they decide what comes next, that could be sufficient to hold serial abusers accountable, regardless of any disbelievers who persist.

The approach is at least worthy of further study. Maybe a pilot program is in order. Let’s test information escrows for sexual harassment—in the U.S. Congress.