Instructors
Lectures
Time: Thursday 12:00–2:00 pm
Location: Calvin 116
Piazza
If you are in the course, you may sign up on Piazza.
Course Description
Human-designed algorithms -- from the digital to the genetic -- reach ever more deeply into our lives, creating alternate and sometimes enhanced manifestations of social and biological processes. In doing so, algorithms yield powerful levers for good and ill amidst a sea of unforeseen consequences. This cross-cutting and interdisciplinary course investigates several aspects of algorithms and their impact on society and law. Specifically, the course connects concepts of proof, verifiability, privacy, security, and trust in computer science with legal concepts of autonomy, consent, governance, and liability, and examines interests at the evolving intersection of DNA technology and the law.
This seminar consists of weekly meetings through the Fall 2018 semester, to be attended simultaneously by faculty, students and scholars based at Harvard, Berkeley, Columbia and Boston University thanks to the IT support of the Berkman Klein Center for Internet & Society at Harvard University. Sessions will be held every Thursday at 3-5pm EST.
Syllabus
Course Assignment
The main assignment in the course will be writing a position paper on a topic related to the material discussed in class. (We will also have a list of suggested topics.) The paper will be written by groups of 2 or 3 students, where each group must contain at least one law student and one CS student.
Students will be expected to work together to write a 5000-7500-word memo that identifies and provides background on the selected issue and outlines solutions. References and citations are essential, and should be included as footnotes.
For cross-institutional groups: the course team will try to assist students to travel in order to collaborate on the class project.
Current Projects
| Team Members | Project Title | Project Abstract |
|---|---|---|
| Sritej Attaluri, Sarah Scheffler | Analysis of Safeguard Implementation Against Governmental Automated Risk-Assessment Mechanisms | In the last decade, advancement and proliferation of big data processing mechanisms have made it easier for governmental entities to enforce standards and requirements on both citizens and non-citizens utilizing large reservoirs of data. This paper seeks to analyze governmental “risk-assessment” systems which use voluntarily and involuntarily obtained data to determine peoples’ rights. Specifically, we’ll be detailing transparency concerns, analyzing various definitions of fairness, and examining multilateral cost-bearing. A variety of case studies will assist with developing a standardized set of potential technical requirements to mitigate abuse and maximize fairness while contextualizing the technologies with their intended purposes. |
| RAYMOND CHU, COLLIN DANIEL CHIN, Kristy Cappelli, Vicky Vlontzou | Antitrust Legislation for Decentralized Consensus - Cryptocurrency Mining and Distributed File Storage | Decentralized platforms such as Bitcoin and IPFS utilize individual computation and hardware contribution to function. The fairness of these platforms relies on the ideal assumption that mining power is distributed equally since all individuals are equally likely to mine a block. However, in reality this is not the case as groups of individuals have agreed to form mining pools such that they have a larger percentage of the overall network and are more likely to receive block rewards. We see that there is an immediate parallel between mining pools and the formation of monopolies within traditional industries as both topics involve the mutual agreement of parties to restrict competition. We wish to explore the implications of antitrust legislation and how it should address applications utilizing decentralized consensus mechanisms, such as cryptocurrency mining pools or distributed file storage. There are many things to consider, such as legitimacy of government interference, liability with creators/users/maintainers, and even the legal designation of decentralized organizations; in these cases, it is also difficult to determine what constitutes legitimate economic or social cause for antitrust legislation to be imposed. As the world tends towards decentralization (in the form of cryptocurrencies, payments, trustless systems, and DAOs), it is important to further understand the social and legal consequences. We wish to explore these controversial topics, establish a systemized way of analyzing decentralized systems from a legal/social perspective, and devise schemes to ensure fairness among all parties. |
| Austin Murdock, Jeongseok Son | Can AI Eavesdrop? | Recently, we have seen a surge in voice assistants, such as Amazon’s Alexa. The deployment of these voice assistants raises serious privacy concerns due to the amount of information they collect and enable companies to infer. In this work, we explore what information Amazon may be able to deduce from the data collection described in the Alexa Terms of Use. We explore laws relevant to Alexa, such as the GDPR, and suggest ways to improve compliance with the law. |
| Malhar Patel, Nisarg Patel | Algorithmic Interpretability in Clinical Settings | With the recent surge of applying deep learning towards clinical diagnostics, there is a lack of interpretable evaluation criteria that incorporates the perspectives of different stakeholders in a medical context. We explore creating this evaluation framework so model vendors, hospitals, patients and insurers are all aligned on providing the best care in a safe manner. |
| Alexis Gauba, Aparna Krishnan | Stable, Scalable, Global Value Transfer | There does not currently exist a way for citizens around the globe to efficiently transfer value cross border with a currency that both scales and maintains stable value. Here, we question how to create such a protocol for a decentralized, scalable, and stable currency, and how this currency can truly be global by being compliant with international law taking into account considerations like GDPR and Adhaar. In order to create efficiency in global processes, there needs to be technological consistency. This is an exploration of how to come to agreement on payments both from theoretical computer science perspective (consensus) and from a human governance perspective (international law). |
| Divya Nekkanti, Julia Schur | Traceable Blockchain Based Evidence vs. The Right to Be Forgotten | A discussion on the growth of blockchain technology compared to the growth of privacy laws and the right to be forgotten. In other words, a discussion of two interactive trends: blockchain (providing a new service which creates a decentralized method of record keeping) versus privacy laws (the right to be forgotten) and how this governs court decisions. Regulated or not, blockchain evidence can be helpful given that it’s a system of “trust” time stamped and validated. We will compare the theoretical acceptance of blockchain technology evidence with how social media evidence has been received in court and the methods under which social media evidence is accepted during trial. (For example, currently social media evidence is gathered through either an affidavit approved by someone (the evidence shows the date, time and content of the post) or using a witness.) We will explore the traces left by blockchain evidence, and the question of if the purchaser/user or the producer of the network is liable if something goes wrong. The paper will also address the regulatory obligations for record keeping. Contrasting EU regulations and US laws as well as the role of regulators to show the need for global agreement because otherwise it won’t work [Indian regulation, Brazil, Asia, EU]. |
| Xinrui Dai | Algorithmic Fairness vs Adversarial Attack | Adversarial examples in machine learning, where the model gets similar inputs but outputs very different labels, are tightly linked to individual fairness and counterfactual fairness. This project aims to build more connections between algorithmic fairness and adversarial attack such as utilizing the adversarial setting to achieve individual fairness. |