Reimagining Settlement with Multi-Party Computation
Alyson Carrel is a Clinical Associate Professor at Northwestern Pritzker School of Law.
Over the past four decades, scholars have debated the impact of private settlement on the broader public good. However, advancing technology in multi-party computation (MPC) presents an option to reimagine how society settles disputes, transforming it from a debate mired in the choice between public or private to a system of pluralistic options, each of which offers parties the choice to settle with varying degrees of both private and public elements. This article uses the cases against Harvey Weinstein and the recent revelations of non-disclosure agreements (NDAs) concealing rampant sexual harassment to demonstrate how a settlement mechanism built on MPC could present a set of pluralistic options that no longer result in the unfortunate choice between prioritizing the interests of individuals versus the needs of society.
The debate about the impact of private settlement is not new. In 1976, Professor Frank Sander spoke at the second Global Pound Conference and advanced the argument that courts moved too slowly, cost too much, and were at times inaccessible to the litigants that needed it most. He described the need for courts to adopt a multi-door approach providing multiple avenues for resolving disputes, including arbitration, mediation, and facilitation. Proponents of these alternative processes argue that not only would it alleviate backlog on court dockets, but that private settlement provides significant social justice benefits. These benefits include i) individual autonomy and self-determination over the outcome of their case, ii) confidentiality to speak freely during settlement discussions and protect individuals from potential reputational harm, iii) creativity and freedom from the limitations of legal remedies, iv) the potential to improve relationships, and v) the ability to have a voice, especially for those that may not have the financial resources to utilize the court system effectively. However, this push for alternatives outside traditional adjudication sparked criticism that expanding the use of private settlement would unnecessarily harm society because it denied the courts its role as protector and reflector of the broader public good. Removing cases from the public court system and placing them in the confines of private settlement distorted society’s understanding of society’s ills, and more generally, placed too much emphasis on the individual resolution and not enough on the society and the systemic impact inherent in a dispute. The resulting debate essentially became one about whose interests to prioritize: the parties, or the broader public good affected by structural harm that the matter represented.
The recent revelations of sexual harassment and assault concealed by Harvey Weinstein and other famous individuals, through private settlement and non-disclosure agreements (NDAs), pointedly demonstrate this debate. Women publicly began to describe previously clandestine incidents of sexual harassment and assault, starting with the Ronan Farrow article in the New Yorker; they named the perpetrators and demanded accountability (of both the perpetrator and the organizations and industries supporting them). This was the first time these women had spoken publicly about the harassment; some women had never reported the harassment in the first place, and others had reported it but agreed to a private settlement that effectively concealed the harassment behind an NDA. The predictions of private settlement’s critics proved prescient: private settlement had prevented individual accountability, failed to change behavior nor protect others from suffering the same harm, and hindered a societal understanding of the larger systemic issues and prevalence of sexual harassment in society. Without formal reporting and knowledge of the perpetrator’s actions, the court could not perform its duty as protector nor reflect and enforce the social and moral norms of society. In this new context, critics once again argued against private settlement and advocated for states and organizations to prohibit the use of NDAs in sexual harassment disputes. They argued that, given these recent revelations, states should prioritize the societal interests of protecting the broader public good over allowing individuals to pursue private settlement, and prohibit the use of NDAs in sexual harassment disputes. They wanted to ensure perpetrators could no longer conceal their harmful behavior from others and protect future third parties from suffering the same harm. Accordingly, California passed such a law prohibiting NDAs. Other states, including Maryland, have passed laws making NDAs in sexual harassment disputes unenforceable, null, or void. Some businesses and employers such as Conde Nast decided to voluntarily declare they would not use NDAs in sexual harassment disputes. Although these new laws and policies limited an individual’s right to choose a private settlement, they did so to ensure bad actors could no longer conceal their behavior.
However well-intentioned, these prohibitions of NDAs disadvantage survivors. As discussed in the debate, there are social justice benefits to private settlement. For some individuals, it is imperative to have the autonomy to settle a claim without going through the costs and perils of going to court. Survivors of sexual harassment often want to deal with the matter quickly and move on. They do not want this experience to follow them in their career and potentially ruin their professional or personal reputation. They often use confidentiality as leverage during settlement negotiations as a bargaining chip with perpetrators who need privacy as well. There is a cruel irony when protecting potential future victims by prohibiting NDAs unintentionally harms the survivors. States recognizing this irony limited the prohibition to the perpetrators by carving out an exception allowing NDAs if the request for the NDA originates with the victim.
Many sexual harassment victims do not report the behavior. They may be afraid they will not be believed or that they will face retaliation if they report. Because of this, some women have relied on more informal and extra-legal mechanisms for reporting sexual harassment and assault, such as the “traditional whisper network,” where they share information among colleagues without ever filing a formal complaint. New mechanisms have developed as well, such as public online forums where individuals post about their experience anonymously (some naming the perpetrator while others do not), and private forums, whether online or offline, where individuals post without anonymity but with the protection of a forum closed to the general public. These extra-legal mechanisms, however, come with limitations. Because they are unofficial, they may not lead to perpetrator accountability. If the identity of the accuser becomes public, the accused may threaten to sue the reporting individual for slander or libel, which may re-victimize the victim regardless of the outcome. Further, individuals accused of sexual harassment in these extra-legal mechanisms voice concern about the injustice inherent in such unregulated sites and the lack of process through which they can defend themselves or ensure their reputation is not unnecessarily damaged.
None of the solutions mentioned above, legal and extra-legal, allow victims to both maintain their privacy while ensuring the public is sufficiently informed about the harm. But the harm is real and so are the calls for solutions that not only protect individuals, but also protect and inform society. The described attempts to develop new mechanisms addressing the need to ensure that NDAs are not concealing ongoing sexual harassment and assault demonstrate a need for a system with pluralistic options that allows individuals who require privacy to feel comfortable reporting while keeping the society informed of the harms. Advancing technology in multi-party computation (MPC) presents the possibility to design such a set of pluralistic options and reimagine settlement.
MPC is a method by which parties can evaluate encrypted secret information to compute results without ever knowing or revealing the underlying data. Owners of the data never relinquish control of their information even though it is used in the computation. MPC scholars often explain the concept using a simple example called Yao’s millionaires’ problem.In this example, two millionaires want to know which one of them has more wealth than the other, while neither of them discloses their wealth. Using MPC, a machine can compute the encrypted data and reveal who has more, without ever revealing, or “knowing” the amount of either millionaire’s wealth. The machine accesses a set of encrypted data, ensuring privacy is maintained while still informing an outcome. Of course, this is an over-simplified example to demonstrate how MPC works. It is not the computation that is remarkable, but the ability to keep the underlying information secure. A trusted third-party could be utilized to house and compute this information, but such information would be susceptible to data breaches or leaks. Because MPC utilizes encrypted data, it is far more reliable in protecting private information. In the sexual harassment context, the question Yao asked – “who has more wealth?” – might become “has this person settled a sexual harassment claim before, and if yes, how many times?” MPC opens up the possibility that parties can settle privately and feel confident their settlement is confidential, while still allowing courts, states, organizations, or society to use that data to compute specific information that could protect the broader public good.
A modified version of MPC has been used on college campuses to provide sexual assault survivors a reporting mechanism for allegations that maintains their privacy unless a subsequent victim alleges an assault by the same perpetrator. This mechanism allows survivors to report the harm without being subjected to the potential for disbelief, retaliation, or reputational damage that might come with publicly reporting an accusation. It also provides a sense of solidarity when a subsequent act is reported and decreases the likelihood that the survivor will be disbelieved. This matching function provides a layer of protection for survivors wary of publicly reporting the harm by allowing them to report it privately as an initial step until they have an ally with which to bring it public. Scholars have proposed a similar step-wise approach to civil sexual harassment settlements that allow individuals to enter into private settlements, with NDAs, if those settlements are “deposited in an information escrow that would be released for investigation by the EEOC if another complaint is received against the same perpetrator.” In other words, individuals maintain their autonomy to choose a private reporting mechanism that only prioritizes the broader public good once there is an indication the perpetrator is a repeat offender. While these proposed systems for sexual harassment cases improve upon the traditional “public” or “private” dichotomy, they still remain problematic. Ian Ayers points out these systems can create a “first-mover advantage,” essentially giving perpetrators a free pass. They also focus exclusively on repeat offenders. Limiting reporting to repeat offenders reinforces the problematic notion that to believe a sexual harassment or assault survivor, there must be multiple allegations. This system does little to inform society about the extent to which sexual harassment and assault is occurring because it is only limited to repeat accounts instead of all accounts.
MPC presents the possibility of creating a much broader reporting system for private settlements not just limited to computing perpetrator matches. A policy could be adopted requiring all private settlements to be maintained and encrypted for computational analysis that opens up the opportunity to leverage the answers to new questions such as: “has this organization settled a sexual harassment claim before, and if yes, how many times?” Or, “how many private settlements dealt with sexual harassment last year?” Or “how many sexual harassment disputes involved an alleged assault?” The possibilities are endless. In her paper “Beyond #MeToo,” Professor Deborah Tuerkheimer argued that survivors need a “wider array of pathways” for reporting sexual harassment. MPC presents the opportunity to develop this array of pathways because it can utilize more than just private settlements and compute more than just information matching.
For instance, instead of prohibiting NDAs, parties could include encrypted information about their private settlement in an MPC mechanism for a court to enforce the private settlement, including any agreed to NDA. This maintains the parties’ autonomy and right to choose settlement but ensures that society can, to varying degrees, evaluate those settlements to ensure the broader public good is protected. Innovations present more questions than answers, and a settlement system built on MPC is no different. Could the parties decide which of the settlement elements to report or would policy dictate the entire settlement be included? Will employers view a reporting system built on MPC as an enhancement over recent laws prohibiting NDAs? How can this system reflect those private settlements in which a court would have never ruled against the employer, but where the company wanted to save money by settling? What will be the appropriate reporting threshold to prevent people from using the mechanism to spread formalized rumors? Will survivors lose settlement leverage if employers know the final settlement must be included in this system and some element of it will be potentially made public? Do parties have to sue before accessing results from the MPC system? Could a settlement system built on MPC change the way society views sexual harassment? The list could go on and on. This article does not have all the answers but simply wants to point out that the technology exists, today, to reimagine settlement.
It is beyond the scope of this short article to identify the entire array available, but here is a brief list of some broad utilizations of MPC. In each example, survivors retain the autonomy to address the harm however they choose, including private settlement, without concealing the impact of these harms from society.
1. Information Escrow for repeat offenders
As already described, MPC is a form of information escrow, in which parties may settle privately unless a subsequent victim files a settlement with the same perpetrator. MPC can provide notice to plaintiff in a subsequent settlement that the defendant has settled a similar claim previously, without disclosing the name of the accuser or precise details of the prior settlement. This example protects the privacy of the individual while still informing future third parties of the settlement and thus the offensive behavior.
2. Information Escrow for repeat organizations defending offenders (even if a single act)
MPC can also be a form of information escrow that informs a plaintiff in a subsequent settlement that the workplace or organization in which the harassment occurred has settled a claim for similar behavior, even if with a different perpetrator. This use of MPC systems expands the value of the information escrow by no longer limiting it to situations in which there is a repeat actor and including situations in which an organization may be harboring or concealing the behavior.
3. Reporting totality of acts in their aggregate
A settlement system built on MPC could also capture information that educates society about this behavior, whether or not the behavior is from a repeat offender. Instead of computing for matches, the system could look at the data in its aggregate to determine the extent of the problem in society and its potential impact on the broader public good. MPC can allow for automatic, cross-jurisdiction aggregations that provide higher-level information about trends in industries. In this way, there is not only accountability for repeat offenders but also anonymized data that can inform the public and raise awareness of widespread behavior, revealing no underlying data about the accuser or the accused. This reporting system mirrors the impact of the informal reporting systems while ensuring the privacy of the individuals.
The recent revelation of individuals using NDAs to conceal rampant sexual harassment and assault demonstrates the importance of the ongoing debate about the impact of private settlement on the broader public good. Current unsatisfactory solutions are limited to a trade-off between whether to prioritize an individual’s interest in privacy over the public’s need for information. MPC presents an opportunity to reimagine settlement, moving away from the old dichotomy of either public or private, to a new set of pluralistic options that allow for both public and private: not only because MPC presents an opportunity to address the use and abuse of NDAs in the sexual harassment setting, but because it presents an opportunity to reimagine how the law interacts with settlement. As critics of private settlement have identified, any dispute between individuals may represent a systemic problem that society has an interest in understanding and protecting future third parties from experiencing. This need cannot come at the expense of the individual’s potential interest in privacy. MPC might allow us to collect data about many cases and run analytics to ensure society has a more accurate understanding of the extent to which certain behaviors are harming the broader public good. With advancing technology like MPC, the law must reexamine the assumptions on which it is based and begin reimagining what is possible, because it is.
 I am grateful to Lynn Cohn, Leonard Riskin, Daniel Gandert, Annalise Buth, Deborah Tuerkheimer, Xiao Wang, and Peter Chan for their support and insightful comments on earlier drafts. I am indebted to Noam Ebner for his leadership in the field and collaboration with me exploring the role of technology in dispute resolution. An earlier version of this article was presented at the Northwestern CS + Law Faculty talk series, organized by Professors Daniel Linna and Kristian Hammond, and the University of Oregon’s Experimental ADR conference, organized and hosted by Professor Jen Reynolds, and I thank all the participants for their comments and suggestions. [back]
 See generally,Owen M. Fiss, Against Settlement, 93 The Yale L.J. 1073 (1984); see generally, Carrie Menkel-Meadow, Whose Dispute Is It Anyway? A Philosophical and Democratic Defense of Settlement (In Some Cases), 83 Geo L.J. 2663 (1995). [back]
 This article approaches the problem from a survivor-centered point of view, but recognizes that settlement is only effective if defendants in civil sexual harassment cases find value in settlement too. [back]
 Frank E.A. Sander, Varieties of Dispute Processing, Address Delivered at the National Conference on the Causes of Popular Dissatisfaction with the Administration of Justice, 70 F.R.D. 79, 111 (1976). For a history of the Global Pound Conferences, see Traum, Lara and Farkas, Brian, The History and Legacy of the Pound Conferences, 18 Cardozo J. Conf. Resol 677 (2017). [back]
 Id. (This is commonly referred to as the “multi-door courthouse”). [back]
 Menkel-Meadow, supra note 2, at 2669-70. [back]
 See, e.g., Nancy Welsh, The Thinning Vision of Self-Determination in Court-Connected Mediation: The Inevitable Price of Institutionalization? 6 Harv. Negot. L. Rev. 1, 15-16 (2001). But see, e.g., Nancy Welsh, Do You Believe in Magic?: Self-Determination and Procedural Justice Meet Inequality in Court-Connected Mediation, 70 SMU L. Rev. 721, 721(2017). [back]
 Confidentiality is now protected to varying degrees by various court rules and law. See, e.g.,Unif. Mediation Act § 8 (2001). But see, e.g.,Christopher Honeyman, Confidential, More or Less: The Reality, and Importance, of Confidentiality is Often Oversold by Mediators and the Profession, Disp. Resol. Mag., Winter 1998, at 12. [back]
 See, e.g., Carrie Menkel-Meadow, Aha? Is Creativity Possible in Legal Problem Solving and Teachable in Legal Education?, 6 Harv. Negot. L. Rev. 97, 97-144 (2001). [back]
 See, e.g., Deborah Thompson Eisenberg, What We Know and Need to Know About Court- Annexed Dispute Resolution, 67 S.C. L. Rev. 245, 248 (2016). [back]
 See, e.g., Nancy A. Welsh, Making Deals in Court‑Connected Mediation: What’s Justice Got To Do With It?, 79 Wash. U. L.Q. 787, 791 (2001); see generally, Nancy A. Welsh, Stepping Back Through the Looking Glass: Real Conversations with Real Disputants About Institutionalized Mediation and Its Value, 19 Ohio St. J. on Disp. Resol. 573 (2004). [back]
 See, e.g., Fiss, supra note 2, at 1075; Laura Nader, Commentary, in THE POUND CONFERENCE: PERSPECTIVES ON JUSTICE IN THE FUTURE 114, 115-19 (A. Leo Levin & Russell R. Wheeler eds., 1979); Harry T. Edwards, Alternative Dispute Resolution: Panacea or Anathema?, 99 Harv. L. Rev. 668, 668-69 (1986); Trina Grillo, The Mediation Alternative: Process Dangers for Women, 100 Yale L.J. 1545, 1561 (1991); David Luban, Settlements and the Erosion of the Public Realm, 83 Geo. L.J. 2619, 2622-26 (1995). [back]
 See generally, Menkel-Meadow, supra note 2, at 2663-71. [back]
 Ronan Farrow, From Aggressive Overtures to Sexual Assault: Harvey Weinstein’s Accusers Tell Their Stories, The New Yorker (Oct. 10, 2017), https://www.newyorker.com/news/news-desk/from-aggressive-overtures-to-sexual-assault-harvey-weinsteins-accusers-tell-their-stories. [back]
 See Fiss, supra note 2, at 1085-87; Luban, supra note 12. [back]
 See Prasad Vasundhara, If Anyone Is Listening, #Metoo: Breaking the Culture of Silence Around Sexual Abuse Through Regulating Non-Disclosure Agreements and Secret Settlements, 59 B.C. L. Rev. 2507, 2520-23 (2018). [back]
 S.B. 820, Cal. 2017-2018 Reg. Sess., 2017 CA S.B. 820(NS), ch.953 (Cal. 2018), https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201720180SB820. [back]
 H.B. 1596, 2018 Reg. Sess. ch.738 (Md. 2018). [back]
 Katie Robertson, Condé Nast to Limit the Use of NDAs, The New York Times (Feb. 21, 2020), https://www.nytimes.com/2020/02/21/business/media/conde-nast-nda.html. [back]
 See generally, Menkel-Meadow, supra note 2. [back]
 See, e.g., Or. Rev. Stat. § 659A.370 (2019). [back]
 Select Task Force on the Study of Harassment in the Workplace: Report of Co-Chairs Chai R. Feldblum & Victoria A. Lipnic, U.S. Equal Emp. Opportunity Comm’n (Jun. 2016), https://www.eeoc.gov/eeoc/task_force/harassment/report.cfm. [back]
 Deborah Tuerkheimer, Beyond #MeToo, 94 N.Y.U. L. Rev. 1146, 1168-74 (2019) (describing a new taxonomy for unofficial reporting in sexual harassment). [back]
 See Tuerkheimer, supra note 26, at 1167-74. [back]
 See Tuerkheimer, supra note 26, at 1188-91. [back]
 Xiao Wang et al., Revisiting Square-Root ORAM, 2016 IEEE Symposium on Security and Privacy, https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7546504. [back]
 Andrew Yao, Protocols for Secure Computations, in Conference Proceedings of the 23th IEEE Symposium on Foundations of Computer Science 160, 160-64 (1982). [back]
 Anjana Rajan et al., Callisto: A Cryptographic Approach to Detect Serial Predators of Sexual Misconduct, Callisto 1, 1 (Nov. 14, 2018), https://www.projectcallisto.org/callisto-cryptographic-approach.pdf. [back]
 Ian Ayres, Targeting Repeat Offender NDAs, Stanford L. Rev. Online (2018), https://review.law.stanford.edu/wp-content/uploads/sites/3/2018/06/71-Stan.-L.-Rev.-Online-Ayres-1.pdf. [back]
 See Tuerkheimer, supra note 26, at 1193. [back]
Alyson Carrel, Reimagining Settlement with Multi-Party Computation, JTIP B