LAW 7688 Research Seminar: The Law of Software
James Grimmelmann
Cornell Tech
Spring 2025
Important Notice
This course is open to both law and non-law students, and to students on both the Ithaca and Cornell Tech campuses. Law students should enroll online, and non-law students should contact me to enroll.
Overview
This is a research seminar on the law of computer software. Specific topics will vary from year to year, but will typically include intellectual property protections for software, constitutional rights to create and run software, embedding legal rules in digital systems, and the regulation of complex artificial intelligence and machine learning systems. The readings will primarily consist of classic and cutting-edge legal scholarship, supplemented with materials on technical background and legal research. Over the course of the semester, participants will research and write a publishable piece of scholarship.
Credits: 3
Meetings: 60 minutes, twice per week, for 13 weeks
Grading: Student option
Satisfies the Writing Requirement: Yes
Course Outcomes
Students who complete this course will be able to:
- Understand how software works and how it is created.
- Understand the competing viewpoints on important issues on the legal treatment of software.
- Have informed and well-supported opinions on topics in software policy.
- Explain technical material clearly.
- Effectively identify the relevant scholarship on a legal issue.
- Efficiently read legal scholarship to identify the important points.
- Research, draft, and edit a work of legal or law-adjacent scholarship.
Who is This Course For?
This course is intended for students who are interested in technology law and want to improve their skills in working with legal scholarship. There are no formal prerequisites, but you should have some familiarity with software, and some familiarity with technology law.
Any of the following is sufficient background in software:
- A college-level course in computer-science.
- Programming experience.
- Significant programming-adjacent work experience (e.g., software project management or web design).
- Multiple CS-adjacent courses dealing with law and/or policy.
- TECH 5300 (may be taken simultaneously).
Any of the following is sufficient background in law:
- A law-school course in computer or Internet law.
- A law-school course in one of the topic areas of the course (e.g., the First Amendment).
- A non-law-school course in computer or Internet policy, if the readings were at least 50% primary legal sources or law-review articles.
- A research project with significant projection onto law (e.g., secure sharing of health data, ML analysis of patent documents, etc.) (may be ongoing).
If you are uncertain whether you have the necessary background, please reach out to me. Everyone’s situation is different.
This course is open to students in all graduate degree programs. Undergraduates will be admitted only in exceptional circumstances.
Policies
Please see the course policies document for information about meeting with me; inclusion; names, titles and pronouns; the history of the site where the course takes place; academic integrity and collaboration; accessibility and accommodations for disabilities; class recordings; professionalism; and health concerns.
Logistics
This syllabus is at http://james.grimmelmann.net/courses/software2025S.
Email: james.grimmelmann@cornell.edu
Huddle: Bloomberg 370
Desk: Bloomberg 3 NW, near the bookshelves
My office hours are whenever I’m free during the workday. You can sign up for a slot at https://jtlg.me/meet. When I’m on campus, we can meet in person in my huddle; when I’m not, there’s a Zoom link on Canvas. If none of the available times work for you, send me an email or DM me on the Cornell Tech Slack.
It’s also always fine just to swing by to see if I’m free. If I have headphones on, just catch my eye. If my huddle door is open, come on in. If it’s closed, it’s closed for a reason (usually a call or a meeting)–send me an email!
Required and Recommended Materials
All required readings will be linked from this syllabus, available online through the through the Cornell library, or posted to Canvas.
I recommend that you have and consult legal citation manual. The nominal standard is The Bluebook, published by a consortium of four law reviews. For law students who intend to practice in the United States, the roughly $50 price tag is a reasonable investment. But for others, you can get by perfectly well with The Indigo Book, a free online reimplementation of the Bluebook’s rules. Introduction to Basic Legal Citation, by Cornell Law’s own beloved former dean Peter Martin, is a highly readable introduction to legal citation that is linked point-by-point to the Indigo Book’s rules.
Although it is not required, due primarily to the unreasonably high price, I also recommend Eugene Volokh, Academic Legal Writing, a writing manual specifically targeted at law-review substance and style.
Class
Most of our sessions will be devoted to careful discussion of a substantive topic (e.g., whether software as such is patentable). The readings for those session will consist of one or more scholarly articles, occasionally supplemented with additional materials for context. These will be interspersed with occasional sessions devoted to the research and writing process (e.g., how to read law-review articles efficiently), which will shift into presentations and discussion of your research as the semester progresses.
I have posted notes and questions for each assignment. These are not afterthoughts. They describe how I plan to attack each article in class discussion. In particular, many of the outlines start with a technical “what is” question, such as “what is encryption?” and we will spend significant class time nailing down these technical concepts. Even if you already know how a technology works, you may be surprised at how much work it takes to make that intuitive understanding precise. We will also pay attention to the descriptions the papers we read use, especially the metaphors they develop. I want you to become as good as the best at explaining technologies.
Each assignment also comes with a list of “Additional Resources.” These are not optional readings; it would be folly to try to read all of them for each class. Instead, you should think of them as starting points if you want to dig deeper on a given class’s topic. Many of them are good enough that I could have built the class around them instead, and all of them are important references if you want to be an expert on their particular topic.
The course will meet in a hybrid format. We will meet in person in Bloomberg room 81 and by Zoom. If you are in the Cornell Tech section, you should join in person unless you are unwell or traveling or have another good reason to join remotely and have confirmed with me in advance.
Attendance in class is required. Especially in view of the other significant demands on your time, I will be understanding about conflicts and flexible in working with you to make alternative arrangements as needed. That said, consistent unexcused absences are not okay, and may lead to a reduced grade or exclusion from the course (after reasonable written warning). Please arrive promptly. I promise that we will end on time, but that means we must start on time. Bring the readings with you, either on your computer or in hard copy.
Assignments
Your work for this class will consist of the following:
First, do the assigned readings and participate in class discussions. I expect all of you to be regular and active participants in the discussions, and to support your classmates in doing so. I understand that everyone has an off day now and then, but this class can only succeed if all of us are fully engaged.
Second, you will write a research paper of at least 10,000 words on a topic of your choosing. I will approve paper proposals on a wide variety of research subjects; the only substantive requirement is that the paper must discuss an issue in which the legal treatment of software depends on the technical details of how that software works.
The paper may conform to the stylistic and scholarly conventions of any relevant academic discipline. For example, law students may write papers in the form of a law-review note, computer-science students may write papers in the form of an ACM conference paper, and so on. You should choose the discipline and form that will be most professionally useful to you. You are not required to submit your papers for publication, but it is my goal for the course that each of you will complete a paper you are proud enough of to want to publish.
The most important goal of this course is to give you practice in writing clearly and precisely about computer technology. Technical readers should find your descriptions are accurate; non-technical readers should find them enlightening. After that, I want to help you have something interesting to say about technology, and be able to explain it persuasively, in a way that other researchers find useful.
Your deliverables for the paper will be on the following schedule:
- Week 2: Preliminary topic proposal
- Week 4: Abstract and preemption check
- Week 6: Bibliography
- Week 8: Detailed outline
- Week 10: First draft
- Week 13: Final paper
- Your paper is due on Tuesday, May 6.
I will meet with you regularly to discuss your projects. I am always available to meet to provide feedback and suggestions, even on short notice.
Grading
Your grades will be determined as follows:
- Class participation: 1/3
- Research paper: 2/3
Schedule
We will usually meet Mondays and Wednesdays 2:15 to 3:15. Our first session will be on January 22, and our final session will be May 5. We will not meet on:
- February 17 (February break)
- March 24 (JG at CS&Law)
- March 26 (JG at CS&Law)
- March 31 (spring break)
- April 2 (spring break)
The following is a rough schedule of topics and readings. I will be filling in more details as we proceed.
-
January 22: Introduction
- Readings: Lawrence Lessig, The Law of the Horse: What Cyberlaw Might Teach, 113 Harvard Law Review 501 (1999)
- Notes: More than any other single work of scholarship, Lawrence Lessig’s The Law of the Horse created and defined the field now known as Internet law. Lessig himself has moved on to copyright policy, Creative Commons, and campaign-finance reform, but Internet law still shows his influence. This article provides a framework for the question this course asks: how does law change when software is involved?
- Questions:
- What kind of argument is Lessig making? It is a technical argument about how the Internet works? A doctrinal argument about how existing law treats the Internet? A policy or normative argument about how law should treat the Internet? A conceptual or jurisprudential argument about how to think about the nature of Internet law?
- If your answer to the first question is “more than one of the above” (hint: it should be), how do the pieces fit together?
- What does the slogan “code is law” mean?
- Why was Lessig’s argument so revolutionary in the late 1990s?
- How well does the article’s theoretical framework hold up? How about its case studies?
- Additional Resources:
- Frank H. Easterbrook, Cyberspace and the Law of the Horse, 1996 U. Chi. Legal Forum
207 (1996). This is the piece that Lessig uses as a jumping-off point. It’s interesting now to look at the substantive portions of Easterbrook’s argument to see how well they hold up.
- Lawrence Lessig, Code: And Other Laws of Cyberspace (Basic Books 1999). This is the book-length version of The Law of the Horse. It was revelatory in 1999. If you want to engage seriously with Lessig’s theory of modalities of regulation, this is essential reading, especially the appendix. Lessig revised Code for a second edition in 2006, but I think it loses some of the focus of the first edition.
- Joel R. Reidenberg, Lex Informatica: The Formulation of Information
Policy Rules through Technology, 76 Texas Law Review 553 (1998). This is the other canonical article that made a similar argument to The Law of the Horse at a similar time.
- James Grimmelmann, Regulation by Software, 114 Yale Law Journal 1719 (2005). This was my student note in law school; it explores Lessig’s metaphor of software as architecture and tries to draw out some general lessons about what software does well and poorly.
- Jonathan Zittrain, The Future of the Internet—And How to Stop It (2008). This is one of a small handful of academic books in technology law to have comparable impact to Code. It is also the one that most directly carries on Lessig’s scholarly approach.
- James Grimmelmann and Paul Ohm, Dr. Generative or: How I Learned to Stop Worrying and Love the iPhone, 69 Maryland Law Review 910 (2010). This review of The Future of the Internet recaps the “architecturalist” tradition of legal scholarship that both Lessig and Zittrain are writing in. Have a look at this if you are wondering about that tradition’s theoretical commitments.
- Bryan H. Choi, The Anonymous Internet, 72 Maryland Law Review 501 (2013). Another response to Zittrain’s book; Choi circles back to the issues of zoning and anonymity from Law of the Horse.
-
January 27: What Is Legal Scholarship?
- Readings:
- Notes: The goal of this class is to understand what makes legal scholarship distinctive from other forms of scholarship, especially in technical fields like computer science.
- Questions:
- What types of scholarship that Minow and Tobin describe strike you as the most interesting and important?
- Where does Lessig’s The Law of the Horse fit in Minow and Tobin’s taxonomy?
- What struck you as most novel or unusual about Lessig’s article as a piece of scholarship, compared with other work you have read? Do the readings for today explain why it had those features?
- Is the law-review submission system as described by Galle functional or dysfunctional?
- Think of a research idea. Would it work as a law-review article? Would it work as an article in a different field? How much do the genre expectations of various academic publication formats drive the substantive content of research?
- Additional Resources:
-
January 29: Software Patents:
- Readings:
- Ben Klemens, Math You Can’t Use: Patents, Copyrights, and Software (2006), chapters 3 and 4 (on Canvas)
- Alice Corp. v. CLS Bank, 573 U.S. 208 (2014). The most important passages are Part I.A (describing the invention at issue) and Part III (the legal analysis).
- Notes: We start the course in earnest with intellectual property (utility patent and copyright) because the question of whether software can be owned forces us to ask what software is. Ben Klemens is a computational social scientist, rather than a lawyer or legal scholar. His argument against software patents is representative of the views of many computer scientists, and also well within the range of mainstream views among legal scholars. Alice is the Supreme Court’s most recent word on the subject.
- Questions:
- What is software, according to Klemens?
- What is the difference between software and hardware? How sharp is the distinction?
- Why does Klemens argue that software can’t software be patented? Is this a technical argument? Doctrinal? Policy? Conceptual?
- Does Justice Thomas’s opinion in Alice adopt Klemens’s reasoning? Reject it? Ignore it? Is the line it draws more or less coherent than the one Klemens draws? Does it have better or worse policy consequences?
- In what sense, if any, is software something that exists in the physical world? What is the significance, or insignificance, of this physicality to Klemens and Thomas?
- Additional Resources:
-
February 3: Software Copyright?
- Readings:
- Notes: CONTU’s recommendations were largely adopted by Congress; the present Copyright Act (as amended) reflects its proposal that software should be protected as a literary work. If there is an orthodox statement of how U.S. copyright law thinks about software, this is it. Richard Stallman (but see) was the driving force behind the creation of the free-software movement, and his essay is a rejection of this orthodoxy. He argues that the entire concept of owning software is a mistake.
- Questions:
- What is software (“computer programs”), according to CONTU?
- Why does CONTU recommend legal protections against the copying of software?
- Why does CONTU recommend treating software as a literary work (the same category used for poetry and novels)?
- Where does Commisioner Hersey (the author of Hiroshima, among many other books) disagree with the CONTU majority? Does his understanding of software differ from the majority’s?
- Where does Stallman disagree with the CONTU majority? Does his understanding of software differ from the majority’s?
- Is Stallman’s argument the same as Klemens’s from last time? Do they rhyme?
- Additional Resources:
- Eben Moglen, Anarchism Triumphant: Free Software and the Death of Copyright, First Monday (Aug. 1999). This is a longer, more detailed, and delightfully sarcastic articulation of the case against software copyright.
- Samir Chopra and Scott D. Dexter, Decoding Liberation: The Promise of Free and Open Source Software (2007). An extended and more academically rigorous exploration of the philosophical case for free software.
- Pamela Samuelson, CONTU Revisited: The Case Against Copyright Protection for Computer Programs in Machine-Readable Form, 1984 Duke L.J. 663. Samuelson’s work towers over the field of software copyright, and I could easily have filled the semester’s reading list with her work. This piece was one of the most significant early critiques of CONTU.
- Pamela Samuelson, Randall Davis, Mitchell D. Kapor, and J.H. Reichman, A Manifesto Concerning the Legal Protection of Computer Programs, 94 Columbia Law Review 2308 (1994). A canonical article on software copyright, combining technical precision with a thoughtful economic analysis. Even decades later, it remains one of the most thorough and careful treatments of the subject.
- David Stein, Hot Apps: Recalibrating IP to Address Online Software, 2024 Wisconsin Law Review 1014. David Stein was a software engineering manager before going to law school and entering academia. This article argues that the shift from apps that run on users’ computers to apps that run as cloud services fundamentally changes the economic argument for software copyright.
-
February 5: Software Copyright
- Readings: Charles Duan, What is Copyrightable in Software? (on Canvas)
- Notes: The previous class was about whether software should be copyrightable at all. But even if it is, there remains a hard question as to which aspects of it are copyrightable. Charles Duan’s article attempts to draw that boundary. Duan (who did his postdoc work at Cornell Tech) was a computer-science major in college and remains an active programmer.
- Questions:
- What is software, according to Duan? How does his description compare to CONTU’s, Hersey’s, and Stallman’s?
- Is Duan’s distinction between the communicative and functional elements of software technically sound?
- Is the distinction consistent with the Copyright Act and caselaw interpreting it?
- Is the distinction consistent with CONTU’s economic arguments for software copyright?
- Additional Resources: There is a vast literature on software copyright, so I can only recommend a few highlights. If you look through the footnotes in Duan’s article, you will find many of the usual suspects. The following are not necessarily the most important pieces, but they are ones that I think are particularly rewarding reads.
-
February 10: AI and Authorship
- Readings: Dan L. Burk, Thirty-Six Views of Copyright Authorship, by Jackson Pollock, 58 Houston Law Review 263 (2020). The late Dan Burk was known for his intellectual tenacity, his personal warmth, and his shrewd sense of humor. All three are on display in this piece. It tackles the question of who (if anyone) should own the copyright to computer-generated works. (This piece is formally inventive. It works here, but this kind of unconventional structure is easy to get wrong. Do not try this at home, unless you are very sure you know what you’re doing.)
- Questions:
- What is a computer-generated work?
- How is a work created using generative AI different, if at all, from a work using analog tools like paintbrushes? From a work generated using digital tools like Adobe Illustrator?
- Do the technical details matter in deciding whether a computer-generated work is copyrightable? Or is it irrelevant how the computer works, because the only important fact is that it is not a human?
- Here is a proposal: computer-generated works are copyrightable, and the copyright is owned by whoever first publishes them. Is this consistent with copyright theory? Would it lead to good results in the real world? What would Burk say?
- This article was written shortly before the recent explosion of generative AI. How well does it hold up?
- Additional Resources:
- Jane Ginsburg and Luke Ali Budiarjo, Authors and Machines, 34 Berkeley Technology Law Journal 343 (2019). This is a long article, but careful and thorough. It covers the same ground as Burk, but more methodically and in much greater depth.
- Pamela Samuelson, Allocating Ownership Rights in Computer-Generated Works, 47 University of Pittsburgh Law Review 1185 (1986). It should be no surprise that Samuelson addressed this question early, or that her work holds up well.
- James Grimmelmann, There’s No Such Thing as a Computer-Authored Work – And It’s a Good Thing, Too, 39 Columbia Journal of Law and the Arts 403 (2016). I argued that “the computer is the author” is a bad answer to the hard and case-specific questions that computer-assisted authorship raises. I still think so, but not as confidently as I did then.
- Ryan Abbott and Elizabeth Rothman, Disrupting Creativity: Copyright Law in the Age of Generative Artificial Intelligence. This is one of the best statements of the case that computers should be treated as authors.
- Bruce Boyden, Emergent Works, 39 Columbia Journal of Law and the Arts 377 (2016). This is the article that made me understand why AI authorship is an intrinsically hard question.
- Katherine Lee, A. Feder Cooper, and James Grimmelmann, Talkin’ ‘Bout AI Generation: Copyright and the Generative-AI Supply Chain, Journal of the Copyright Society of the USA (forthcoming). A more recent take. The computer-authorship material is in Part II.A, but there is also a detailed description of how modern generative-AI systems work in Part I, which you may find useful.
- Carys Craig and Ian Kerr, The Death of the AI Author, 52 Ottawa Law Review 33 (2021). A very different take on AI authorship, one that focuses on the social role of “author” rather than the mental and physical processes that results in a work.
-
February 12: Legal Citation
- Readings:
- Notes: In some ways, legal citation is a citation system like any other–a set of mechanical rules for how to describe a source. But it has some unusual properties that are worth paying attention to. First, it is a more elaborate system than any other I am aware of, with a variety of unique conventions for what information must be included and how it should be presented. Second, because it is used by practicing lawyers as well as scholars, and because the nature of authority in law is different than in other fields, it serves some goals (such as rapidly indicating why a source is being cited) that other citation systems do not. Peter Martin is a former dean of Cornell Law School and deeply knowledgable about the history and practice of legal citation.
- Questions:
- How is legal citation different from citation in other fields?
- What are the goals of a citation system? How well do the citation systems you have used meet them?
- Regardless of how citations are formatted, why does legal scholarship use so many of them?
- Additional Resources:
- Anatomy of Basic Legal Citations. A nicely-formatted cheat sheet for the most common types of legal citations.
- Orin S. Kerr, A Theory of Law, 16 Green Bag 2d 111 (2012). This is a joke, but also maybe not quite a joke.
- There is a small cottage industry of articles attacking the Bluebook. Some call for reforming it to be less detailed, some call for throwing the whole thing out. Notable examples include Paul Gowder, An Old-Fashioned Bluebook Burning , 1 Northwestern Law Journal Des Refusés 1 (2024); Richard A. Posner, Goodbye to the Bluebook, 53 University of Chicago Law Review 1343 (1986); James Ming Chen, Something Old, Something New, Something Borrowed, Something Blue, 58 University of Chicago Law Review 1527 (1991); David Ziff, The Worst System of Citation Except for All the Others, 66 Journal of Legal Education 668 (2017)
- For background on the history and evolution of the Bluebook and legal citation, see Fred R. Shapiro and Julie Graves Krishnaswami, The Secret History of the Bluebook, 100 Minnesota Law Review 1563 (2016)
- Peter Martin’s Citing Legally blog features some outstanding deep dives on the nuances of legal citation.
- There have been numerous attempts to automate legal citation. Many of them are badly incomplete, make basic and easily spotted mistakes, or both. The only two software packages that I can recommend are Juris-M (a Zotero fork with a Word plugin) and Hereinafter (a LaTeX package). I have used both, currently use Hereinafter, and can provide pointers and advice on getting started if you would like.
-
February 19: AI and Infringement
- Readings: Benjamin L.W. Sobel, Artificial Intelligence’s Fair Use Crisis, 41 Columbia Journal of Law and the Arts 45 (2017).
- Notes: Ben Sobel is a post-doctoral fellow at Cornell Tech, and he will join us for the discussion. This is not a comprehensive overview of the infringement issues involved in AI training; instead, it focuses on one of the most important ones.
- Questions:
- How are (potentially) copyrighted works used in AI training?
- Why would copyright law have a problem with this?
- When should fair use protect AI training? If your answer is not “always” or “never,” how should courts draw the line?
- This article was written before the recent explosion of interest in generative AI. Does anything in the analysis change when AI is used to create new creative works, rather than to make (non-expressive) decisions?
- Additional Resources:
- Amanda Levendowski, How Copyright Law Can Fix Artificial Intelligence’s Implicit Bias Problem, 93 Washington Law Review 579 (2018). Another early article about copyright and AI training, but one that framed the problem a little differently than Sobel did: copyright stands in the way not of AI innovation but of AI fairness. Compare Amanda Levendowski, Resisting Face Surveillance with Copyright Law, 100 North Carolina Law Review 1015 (2022), in which Levendowski argues that copyright should be used to limit some forms of AI training in the name of other important social values.
- Mark A. Lemley and Bryan Casey, Fair Learning, 99 Texas Law Review 743 (2021). A particularly clear and well-written take on fair use and AI training.
- James Grimmelmann, Copyright for Literate Robots, 101 Iowa Law review 657 (2016). My own piece on bulk technological uses of copyrighted works applies to AI, but is not entirely about AI. It’s a little more polemical than the other readings here.
- Benjamin L.W. Sobel, Elements of Style: Copyright, Similarity, and Generative AI, 38 Harvard Journal of Law and Technology (forthcoming). A more recent piece by Ben, which focuses on substantial similarity of AI outputs rather than on fair use for AI training.
- Mark A. Lemley, How Generative AI Turns Copyright Upside Down, 25 Columbia Journal of Law and the Arts 21 (2024). Another Lemley piece, one that gets at the subtle and tricky relationship between AI prompts and AI outputs.
- A. Feder Cooper and James Grimmelmann, The Files are in the Computer: On Copyright, Memorization, and Generative AI, Chicago-Kent Law Review (forthcoming). Is a model a “copy” of the works it was trained on? We think the answer can be “yes,” at least sometimes—when a generative model has memorized a work, in the sense that it can produce a substantially similar version of that work as an output.
- Matthew Sag, Copyright Safety for Generative AI, 61 Houston Law Review 295 (2023). One of the most important articles on fair use and AI training from the modern era, i.e., after the launch of ChatGPT. Sag makes several excellent points about the different ways in which an AI system can learn from a copyrighted work, some of which may be infringing and some of which may not.
-
February 24: AI and Defamation
- Readings: Eugene Volokh, Large Libel Models? Liability for AI Output, 3 Journal of Free Speech Law 489 (2003).
- Notes: Eugene Volokh is one of the preeminent First Amendment scholars working today, and he has a long-standing interest in technology and Internet law, dating back to his extensive experience as a programmer. Note that this is a big and slightly sprawling article; many doctrinal articles are, because they attempt to cover all of the relevant issues that could arise in a lawsuit. (I’ve certainly been guilty of this myself!) As you read, try to pin down which issues are fundamental and raise genuinely new challenges, and which are likely to become non-issues over time.
- Questions:
- Should we treat the outputs of AIs as potentially defamatory? Aren’t they obviously the results of purely digital processes, with no semantic meaning?
- Relatedly, why should anyone take what ChatGPT outputs seriously? Should the legal system presume that everyone should know that generative-AI chatbots hallucinate all the time?
- Can an AI system have actual malice? What would it take to conclude that ChatGPT emitted an output “with knowledge that it was false or with reckless disregard of whether it was false or not”?
- Should OpenAI have a duty to prevent ChatGPT from hallucinating defamatory lies? Should it have a duty to investigate and fix ChatGPT after someone points out a defamatory lie that it emitted?
- Additional Resources:
- Another doctrinal analysis is Leslie Y. Garfield Tenzer, Defamation in the Age of Artificial Intelligence, 80 NYU Annual Survey of American Law 135 (2024).
- Toni M. Massaro and Helen Norton, Siri-ously? Free Speech Rights and Artificial Intelligence, 110 Northwestern University Law Review 1169 (2016), and Toni M. Massaro, Helen Norton, and Margot E. Kaminski, Siri-ously 2.0: What Artificial Intelligence Reveals About the First Amendment, 101 Minnesota Law Review 2481 (2017) are listener-oriented takes on AI speech, written before ChatGPT upended everything.
- Dan Burk, Asemic Defamation, or, the Death of the AI Speaker, 22 First Amendment Law Review 189 (2024) is the most pungent argument that AI outputs aren’t even meaningful enough to be defamatory. Another take arguing that AIs lack the intentions required in law for certain kinds of liability is Ian Ayres and Jack M. Balkin, The Law of AI is the Law of Risky Agents without Intentions, University of Chicago Law Review Online (forthcoming).
- James Grimmelmann, The Defamation Machine (38th Annual Silha Lecture 2023) is my take on whether AI outputs have meaning and whether AI can have actual malice. It has pictures!
- Lawrence B. Solum, Artificial Meaning, 89 Washington Law Review 69 (2014), is a philosophically sophisticated take on where the meaning, if any, in AI outputs comes from.
- Peter Henderson, Tatsunori Hasimoto, and Mark Lemley, Where’s the Liability in Harmful AI Speech?, 3 Journal of Free Speech Law 589 (2023). Another recent paper, from a mixed CS+law team of authors.
- Stuart Minor Benjamin, Algorithms and Speech, 161 University of Pennsylvania Law Review 1445 (2013). An older but very careful analysis of whether and when algorithmic outputs can be “speech” covered by the First Amendment. Compare Tim Wu, Machine Speech, 161 Pennsylvania Law Review 1495 (2013), from the same symposium, which takes a more generally skeptical view. James Grimmelmann, Speech In, Speech Out, in Ronald K.L. Collins & David M. Skover, Robotica: Speech Rights and Artificial Intelligence 85 (2018), is my take on some of the issues.
-
February 26: Layering
- Readings: Lawrence B. Solum and Minn Chung, The Layers Principle: Internet Achitecture and the Law, 79 Notre Dame Law Review 815 (2004)
- Notes: This article is absurdly long, and I don’t expect you to read every part of it closely (or necessarily at all). Instead, the article is interesting because it is an attempt to make a technical principle normative. Solum and Chung argue that the architecture of the Internet as it currently exists (as of 2004, when they were writing, that is) is layered. They then claim that laws should respect that architecture: lawmakers should not write laws that cross between layers, or that undermine the layering itself. This is an interesting kind of argument! Remember that Lessig said law can change architecture–Solum and Chung are saying that architecture should drive law.
- Questions:
- What is “layering”? Is it something that we only see on the Internet, or can other systems also be layered?
- Was the Internet actually layered in 2004 when they wrote? Is it layered now in 2023?
- Just because the Internet is layered now, does it follow that it it should be layered? (What do Solum and Chung say? Are there counterarguments? Better arguments that they fail to make?)
- What makes a law “layer-crossing”? Is this a coherent category, or a catch-all term for bad Internet laws?
- Does the layers principle imply network neutrality?
- Have you seen this type of attempt to make technical principles normative anywhere else? Are those other attempts persuasive?
- Additional Resources:
- The Cursed Computer Iceberg Meme. Computers are deeply weird. We talk about them using logical and well-structured abstractions, but sometimes—often—those abstractions break down. This is a compilation of incredible, and frequently hilarious, stories about times when those abstractions broke down.
- Eric Wustrow, Scott Wolchok, Ian Goldberg, and J. Alex Halderman, Telex: Anticensorship in the Network Infrastructure, Proceedings of the 20th USENIX Security Symposium (USENIX Security ‘11) (2011). An interesting technical proposal to help users circumvent governmental censorship, which depends in a fundamental way on violating the layers principle. See also the authors’ companion sites for Telex and its descendant Refraction Networking and my blog post about Telex, Planet Telex.
- Larry Patterson and Bruce Davie Computer Networks: A Systems Approach (Morgan Kaufmann 6th ed. 2021). This is my personal favorite networking textbook; it presents the standard model of the Internet. It is useful to read while pondering the question of where the layers principle comes from.
- Pamela Zave and Jennifer Rexford, The Real Internet Architecture (Princeton University Press 2024). This book gives a more modern model of Internet architecture, including developments like VPNs, firewalls, and CDNs. It is useful to read while pondering whether the Internet today obeys the layers principle.
- Barbara van Schewick, Internet Architecture and Innovation (MIT Press 2010). A thorough and detailed analysis of the role that the layers principle and the related end-to-end principle play in Internet policy. The literature review is exceptional; if you are going to do serious research on Internet architecture and law, this is an essential reference.
- J[erome] H. Saltzer, D[avid] P. Reed, and D[avid] D. Clark, End-to-End Arguments in System Design, 2 ACM Transactions on Computer Systems 277 (1984). This is the paper that introduced the end-to-end principle; it remains highly readable and a canonical reference.
- Jerome H. Saltzer and M. Frans Kaashoek, Principles of Computer System Design: An Introduction (Morgan-Kaufmann 2009). This is a full-on computer-science textbook about system design. Part I is only legally available in print, but Part II is free online and includes a chapter about layers and modularity in networks. A good place to look to understand how system designers think about abstractions.
-
March 3: Protocols
- Readings: Eric J. Feigin, Architecture of Consent: Internet Protocols and Their Legal Implications, 56 Stanford Law Review 901 (2004)
- Notes: This is another paper on the legal consequences of technical principles, published in the same year as The Layers Principle. But where Solum and Chung focus on the consequences for regulators of the fact that the Internet is designed in a particular way, Feigin focuses on the consequences for private parties and judges. Given how Internet protocols work, Feigin argues, people who use those protocols in particular ways should be treated as having consented (or not) to particular conduct. This is also an interesting kind of argument, but be clear that it is a different kind of argument than we discussed last time.
- Questions:
- What is a protocol?
- Was Feigin’s description of Internet protocols accurate in 2004? Is it accurate now in 2023?
- Is Feigin right that use of an Internet protocol is a kind of consent? If so, is it the kind of consent that can be withdrawn by an explicit statement to the contrary?
- What kind of technical information do you need about a protocol to attribute legal consequences to its use? What else do you need to know?
- Could someone define a new protocol that is like IP or TCP or HTTP but which does not have these consent-granting features?
- Additional Resources:
-
March 5: Workshop on writing a clear abstract and introduction
- Readings: Randall Munroe, Up Goer Five, XKCD.
- Notes: For class today, write the first part of your paper using only the ten hundred words people use most often. It may help to use this box you can type in to check your work.
- Additional Resources:
- March 10: Smart Contracts
- March 12: Smart Contracts
- March 17: Software Liability
- March 19: Software Liability
- April 7: The Fifth Amendment
- April 9: t/b/d
- April 14: t/b/d
- April 16: t/b/d
- April 21: t/b/d
- April 23: The Law Review Publication Process
- April 28: Paper Presentations
- April 30: Paper Presentations
- May 5: Paper Presentations