News & Events
  /  
Seminars, Workshops, & Talks
CS + Law Faculty Talks

Wednesday, February 26, 2020
5 – 8 p.m.


Pritzker School of Law, RB 140

Reception including refreshments and appetizers following the event

The Computer Science + Law Faculty talks will bring together faculty interested in the intersections of computer science and law with the purpose of fostering collaboration and interdisciplinary research.

Register for the event

Agenda

5 - 5:02 p.m. — Welcome by Samir Khuller, Kris Hammond, and Dan Linna
5:02 p.m. - 5:06 — Kris Hammond: Northwestern "AI and Law" proposal
5:07 - 5:10 — Dan Linna and Samir Khuller: Current and future activities

Short talks by CS + Law faculty

Jump to a Section

 

Technology, Law & Society

Jim Speta, Elizabeth Froehling Horner Professor of Law and Vice Dean
Internet Platforms and the Public Utility Impulse
Both those seeking and those abhoring greater government regulation of Internet platforms appeal to the history of public utility regulation, often taking a normative perspective that such regulation has a single public interest story to be told. From a positive perspective, one can clearly identify the elements of public utility regulation. Doing so reveals the inconsistencies (and worse) of the current debate, for almost no regimes proposed or criticized include all of its elements. From the perspective of motive or justifications, moreover, public utlity regulation was not imposed for consistent, economically-identifiable reasons, and suggestions that it was largely track deregulatory ideologies. Being honest about its historical elements allows us to see a more probable path forward, one which more resembles comprehensive health and safety regulation than the control of utilities.

Juliet Sorensen, Clinical Professor of Law
Hack for Social Good: The Intersection of Human Rights, Public Health, and Technology in Lagos
Since 2016, the Northwestern Access to Health Project has trained and supported community health educators and advocates ("CHEs") in the slums and informal settlements of Lagos. Since 2018, Access to Health has partnered with Slalom Consulting to design a website and mobile app for the CHEs as a ready reference that will also serve as means to gather data. This law-tech initiative from the ground up continues to be an iterative collaboration as we refine it with input from our partners in Lagos and also consider key objectives of sustainability and capacity-building.

Paul Gowder, Professor and O.K. Patton Fellow in Law (Paul will join the Northwestern Law faculty in the fall of 2020)
Platforms and Policies
Gowder argues that the best explanation of the increasingly legalistic approach exhibited by major internet platforms (especially, but not exclusively, social media, as identified by folks like Klonick) is that the behavior management problems of such platforms are functionally similar to problems confronted by polities at scales ranging from cities to nation-states. But this is an opportunity: it allows us to bring in generations of research in, inter alia, institutional political science, organizational economics, and constitutional theory, to help platforms figure out their behavioral control problems. Examples, among many, include the strategic literature on the rule of law and the Hayek/James C. Scott line of work about the knowledge problems of centralized control.

Xiao Wang, Assistant Professor of Computer Science
From Brandeis to Brandeis
The Right to Privacy was introduced back in 1890 by Warren and Brandeis. It fundamentally shapes our expectations of privacy. Now, we are still living in a world distant from what Warren and Brandeis pictured for us back in 130 years ago. In this talk, Wang will talk about essential tools towards this goal.

Laura Nirider, Clinical Associate Professor of Law
Lawyering in the Age of Social Media
In 2015 and 2018, tens of millions around the globe learned about the Center on Wrongful Convictions' client, Brendan Dassey, on the two-season hit Netflix series Making a Murderer. Countless ordinary people were moved by Brendan's story of wrongful conviction — and took to social media to express their support. Twitter, Reddit, and Instagram now host "stan" followings and "free Brendan" activist groups that number in the tens or even hundreds of thousands. But what does a lawyer do when so many people around the globe want to support her client — including in ways that she cannot control? Quite by accident, the CWC's faculty and students have been at the forefront of lawyering in the age of social media. We've been among the first to develop advocacy techniques designed to actively build, influence, and deploy global trends on social media to the advantage of our client, while staying within traditional ethical boundaries. But lawyering in this social media moment raises important questions that will arise in every high-profile case going forward: How should advocates interact with social media's profound ability to influence decision-makers either for — or against — one's client? Do traditional lawyering principles demand that lawyers ignore social media and its influence? Or does ignoring social media come at too great a cost to one's client? More profoundly, can social media make room for a participatory vision of justice — one that allows previously unheard voices to resound — while stopping short of justice by Twitter mob?

Nick Diakopoulos, Assistant Professor
Synthetic Media and Free Speech
New policy and legal questions are emerging at the intersection of speech and the capabilities of AI to both generate and moderate that speech. For instance, rule-making processes might be overrun by individuals using synthetic media generation techniques (i.e. deepfake text) to produce millions of different versions of their comment. Social media may be swamped by political actors using the same technologies to push propaganda and undermine legitimate debate. Should bots be able to interact with humans online without disclosure? Should they be afforded the same speech rights given their ability to crowd other speech and undermine the value of counter-speech? Moreover, how can free speech be maintained at the scale of the largest online communications platforms while using AI to help moderate and screen harmful (e.g. hate) speech? Statistical moderation tools necessarily have an envelope of uncertainty and struggle with intent, creating challenges around preserving legitimate political speech such as satire and parody.

Text Analytics and Interpretation

Sarah Lawsky, Benjamin Mazur Summer Research Professor of Law
Teaching Algorithms and Algorithms for Teaching
Much of the complexity and difficulty of statutory law comes from the large number of rules involved, and from understanding the interaction of rules whose language is relatively clear. Lawsky will present a program she has written that generates tax practice problems for students by combining various complex but unambiguous rules. In addition to being a useful teaching tool, the program demonstrates potential benefits of rules-based programming of statutory law.

Chris Riesbeck, Associate Professor of Computer Science
Understanding Language and Legal Texts
This non-technical talk will contrast processing texts (smart search and extraction) with understanding them, give everyday examples to illustrate how language understanding is not about the meanings of words but the recognition of prior knowledge, and describe past work using this model. Discussion could focus how understanding would be enable intelligent use of documents such as contracts and wills.

Daniel Rodriguez, Harold Washington Professor of Law
Statutory Information as Data: ML and Legislative Interpretation
Debates about judicial and agency interpretation of ambiguous statutes often turn on matters of interpreter competence and the availability of relevant information. These debates exist alongside the hoary debate about the proper role of interpreters in our constitutional system. Advances in machine learning, and, as well, the increasing availability of data concerning legislative history and administrative agency performance in implementing statutes illuminate (at least) the questions at the heart of the first debate. This short presentation sets the table of this issue and, more ambitiously, provides some preliminary thoughts about the connection between textual ambiguity in the law, computation and big data.

Emerson Tiller, J. Landis Martin Professor of Law & Business
Textual Analysis and the Law
There is growing interest by legal academics in the use of textual analysis for studying legal process and decision-making. A conference titled "Law and Textual Analysis" was organized and held at Northwestern Pritkzer Law School in April 2019 (funded by Kellogg's Ford Center for Global Citizenship and Pritzker Law School). This presentation reflects on the various topics and tools used in the papers presented by legal academics at this conference.

BREAK: 6:30 - 6:45 p.m.

Law and Data Analytics

John McGinnis, George C. Dix Professor in Constitutional Law , with Steven Wasick
Law's Algorithm
Law fundamentally concerns information — providing information to the community about the content of legal norms and, eliciting information about the world from the disputes before a court. Information theory helps us understand how the computer — the key technology of our day — is changing legal search and thereby the efficient form of law. Information theory focuses on the signal to noise ratio of communication. The key to progress in creating a better computerized legal search engine is to reduce the signal to noise ratio in the link between the user and the search engine. The ongoing improvement in legal search modifies the optimal form of the law by changing the cost of finding it. It rebalances the weights in the classic debate between rules and standards, because exponential increases in computational power make standards relatively more attractive than rules by decreasing the costs of their application.

Larry Birnbaum, Professor of Computer Science
What We Can Tell About You From Twitter
It's now widely understood that it's possible to infer many facts — and specifically, many facts that one might prefer to keep private — about people from their online activity, in particular their activity on social media. What might not be so widely understood is just how easy and obvious this technology now is to implement. In this talk, Birnbaum will demonstrate some systems for predicting preferences, in particular voting preferences, based on Twitter data. Accuracy of these systems ranges from 85-90%. The latest of these is a platform for automatically constructing such systems, built by three undergraduate students at Northwestern in two quarters.

Alyson Carrel, Clinical Associate Professor of Law
Impact of Legal Analytics on Settlement
With the rise of data-driven lawyering and online dispute resolution, data analytics will have an increasing impact on traditional dispute resolution and confidential settlements. Lawyers are beginning to recognize the value of using legal analytics to predict the outcome of litigation and inform settlement decisions. This data typically stems from court documents and litigation outcomes. But given the proliferation of online dispute resolution in courts and the potential for increased collection of data related to settlement behaviors and outcomes, a new data set may emerge that can address debated issues in the ADR field such as the value of a BATNA analysis in the age of the diminishing trial or the harm to public policy and marginalized populations caused by confidential settlements and processes.

David Schwartz, Professor of Law
Testing the Efficacy and Efficiency of In-House Lawyers
This Article seeks to empirically test whether in-house counsel perform as high quality of legal work as outside law firm lawyers and do so in as cost-efficient a manner. To empirically compare the efficiency and effectiveness of in-house lawyers to outside lawyers, the Article examines the process and result of getting a patent from the United States Patent and Trademark Office (USPTO)—a process called patent prosecution. The study observes whether the in-house or outside nature of the attorney handling the prosecution correlates with recognized measurements of patent quality and efficient and compact prosecution. This data allows us to test whether certain assumptions about the economic benefits of in-house counsel compared to outside counsel are true, particularly in the transactional law context with highly granular data—the administrative patent prosecution history and resulting patent.

Laura Pedraza-Farina, Associate Professor of Law
The Case for Tailoring Patent Strength to Innovation Diffusion
The patent system is often justified on the basis of two theories. The first, incentive theory, views patents as incentives to invest in innovation. The second, disclosure theory, views patents as incentives to disclose information about an invention. Both justifications are tied to ex ante behavior: incentives to invent foster innovation up to the time of patent filing, and disclosure is achieved once the patent document is made public, typically 18 months after filing. This Article argues that this ex ante view of patentees’ behavior is incomplete both as a matter of theory and as a matter of actual doctrinal practice. It makes the theoretical case that ex post patent strength should be tied to efforts by the patentee to diffuse his or her innovation. And it shows how understanding patent rights as tools to facilitate the diffusion of technological knowledge helps explain otherwise puzzling doctrinal decisions.

Technology and Teams

Jason Hartline, Professor of Computer Science
Designing Incentives (With application to peer grading)
A syllabus organizes the work that happens — by teachers, course staff, and students — in a course by establishing incentives for individuals and indirectly governs the outcomes that are obtained. This talk investigates this perspective on peer grading where student work can potentially be graded faster and more accurately, and students learn in the process, when grading is done by computer facilitated and aggregated peer grading.

Samir Khuller, Peter and Adrienne Barris Chair and Professor of Computer Science
On Building Diverse Teams
Matching is a popular method for assigning people to teams, optimizing some utility function. However, sometimes teams may benefit from being diverse. How does one model diversity and the utilize it in team formation? We discuss some basic questions in this framework.

Evaluating Legal Technology

Maria Amparo Grau Ruiz, Visiting Professor of Law
The Attempts to Regulate the Social Impact of Robotics and AI in the EU
The European Institutions are considering how to regulate robotics and AI, and to which extent. The EU citizens are concerned about their rights. The research on Ethical, Legal and Socio-Economic aspects is still developing, and increasing attention is being paid to them in technical circles (e.g. ERF). There are important efforts to find the best ways to implement Responsible Research and Innovation principles. Several international teams across Europe are working with the necessary interdisciplinary approach. INBOTS H2020 project will deliver the White Paper for the Regulation of Interactive Robotics at the end of this year. We are now focusing on human rights (specially, privacy) and sustainability (particularly, the environmental perspective). However, the solutions cannot be only regional and we should aim to establish certain common global standards. Strategically, it would be advisable to boost this movement at the highest institutional level with the active involvement of Bar Associations, and the actors in each field (e.g. the International Fiscal Association will devote its 2021 Congress to Big Data and Tax).

Kris Hammond, Bill and Cathy Osborn Professor of Computer Science
Viewing Technology Through the Lens of Its Application
The technologies of intelligence on the rise and we are being called on to evaluate their impact. The trend has been to focus on the technologies themselves and to think in terms of restricting or allowing wholesale. Technologies such as facial recognition or prediction are viewed as black boxes that are either going to be unrestricted or banned outright. In this talk, Hammond will outline a different approach in which we consider the particular applications and their impact in different situations and walk through an approach to regulation that focuses on this application layer rather than the technologies that support it.

Dan Linna, Senior Lecturer and Director of Law and Technology Initiatives
Developing Standards for the Evaluation of Legal Technology
How can we ensure that legal technology helps and does not harm people? We hear a variant of this question with each launch of a technology tool to improve access to legal services. As several jurisdictions move towards relaxing rules limiting who can deliver legal services, there is an urgent need to develop objective standards to evaluate legal technology. Large law firms, corporate legal departments, and other legal technology developers and users also need to evaluate these technologies. Evaluation of information retrieval systems and eDiscovery gives us a starting point, but much more work needs to be done.

Additional speakers will be announced.

Organizers

Samir Khuller
Peter and Adrienne Barris Chair and Professor of Computer Science

Dan Linna
Senior Lecturer and Director of Law and Technology Initiatives

Jeff Stiers
Senior Director of Strategic Projects