Three Laws Blog: Law and Artificial Intelligence

Bartholomew on A Right to Be Left Dead @markbartholo @UBSchoolofLaw @CalifLRev

Mark Batholomew, SUNY Buffalo Law School, is publishing A Right to Be Left Dead in the California Law Review. Here is the abstract.

Technology forces us to contemplate a counterpart to the right of privacy—Brandeis and Warren’s “right to be left alone”—for the age of AI: the right to be left dead. Traditionally, it has been presumed that even if Brandeis and Warren’s right constitutes “the most comprehensive of rights and the right most valued by civilized men,” it does not apply to those no longer alive. The question is whether we need a new calculus at a time when anyone can be reanimated and made to behave in a manner indistinguishable from their living presence. This Article interrogates the need for a right to be left dead and takes some preliminary steps towards defining its contours, chief among them an awareness that an individual right to prevent unauthorized reanimations of the dead must look very different than the existing privacy, consumer protection, and property laws marshalled against unauthorized invocations of the living. A new right to be left dead can serve important social goals, but it must be narrowly construed, yielding to the critical interests of the living when necessary. To the extent legislatures and courts are beginning to grapple with the nascent problem of digital reanimation, they have failed to account for the most fundamental dividing line in existence—the line between life and death.

Download the article from SSRN at the link.

Heidi Härkönen on The Impact of AI on the Fashion Section: A Moral Rights' Perspective @FashionLawHeidi @UniTurkuLaw @eLAWnora @Ire_Calboli @routledgebooks

Heidi Härkönen, University of Turku Faculty of Law, is publishing The Impact of Artificial Intelligence on the Fashion Sector: A Moral Rights’ Perspective in Routledge Handbook of Fashion Law (E Rosati and I Calboli, eds. Routledge, 2024). Here is the abstract.

This chapter discusses the copyright-related impacts that sophisticated artificial intelligence (AI) has on authors in the fashion sector. The increasing use of AI in fashion design may threaten the position of human designers, as AI is cost-efficient and, unlike human designers, is not entitled to IP rights over its ‘creations’. The fashion sector should be concerned about AI’s effects on human designers because they are already in a vulnerable position in the fashion value chain. From a copyright standpoint, the practices of the fashion industry in relation to the treatment of (human) designers are questionable. This is particularly visible in the way designers’ moral rights are systematically ignored, diluting designers’ authorial role and transparency in terms of the source of designs. At the same time, fashion companies’ increasing use of AI raises further transparency issues regarding the origin of designs, causing uncertainty both for consumers and rivals. Increasing attention to designers’ moral rights, in particular the right of attribution, could help to improve human designers’ position, balance the unjust status quo, and increase transparency. Moreover, a functional equivalent of the right of attribution to AI-generated output would improve transparency in fashion and creative sectors in general.

Download the chapter from SSRN at the link.

Murray on AI for Academic Support in Law Schools and Universities @ukcollegeoflaw

Michael D. Murray, University of Kentucky College of Law, has published Artificial Intelligence for Academic Support in Law Schools and Universities. Here is the abstract.

The current models of verbal generative artificial intelligence (AI)—Bing Chat, GPT-4 and Chat GPT, Bard, Claude, and others, and the current models of visual generative AI—DALL-E 2, Midjourney, Stable Diffusion, and others—can play a significant role in academic support in law schools and universities. Generative AI can help a student learn and understand material better, more deeply, and notably faster than traditional means of reading, rereading, notetaking, and outlining. AI can explain, elaborate on, and summarize course material. It can write and administer formative assessments, and, if desired, it can write self-guided summative evaluations and grade them. AI can translate material into and from foreign languages with a fidelity to context, usage, and nuances of meaning not previously seen in machine learning or neural network translation services. AI also can visualize material using the tools of visual generative AI that literally paint pictures of the subjects and situations in the material that can overcome students’ literacy issues both in the native language of the communication and in the students’ own native languages. Beyond supporting student learning and academic success, AI can be a democratizing force because it can empower students to begin writing or drawing or painting at a level that their own life experiences and education have not prepared them or enabled them to participate in. AI can empower students to perform creative, artistic, or literary activities related to legal education and law practice at a high level, catching them up to where other classmates would start. First-generation college-goers and graduate students can use the collective knowledge of a large language model to bring themselves to a higher starting point in the process of gaining admission to and finding success in legal education and ultimate in the practice of law. The current text-based generative AI models in the form of Microsoft’s BingChat, Goggle’s Bard, OpenAI’s ChatGPT and traditional methods of tutoring and academic support. It does so through the multimodal nature of its skills: AI can explain, elaborate on, and summarize course material; it can interpret, translate, visualize, or reorder parts of the material. AI can evaluate and correct the grammar, spelling, syntax, and style of a piece of student writing, a task which campus writing centers often avoid for pedagogical reasons or simply logistical and resource-driven reasons. AI has become a master translator moving easily from communications in one language and converting them into many languages, and at the same time monitor the grammar, spelling, syntax, and style of the translated work for fidelity of usage in the target language. At the farther reaches, AI can communicate to illiterate and less than fully literate students because several of the current AIs or their close corporate cousins speak the language of images (i.e., visual communication) by generating visuals to illustrate, depict, diagram, or graph a concept. AIs can deliver the gold-standard level of one-on-one, personalized attention for tutoring and support. Naturally, with this amount of power placed in the hands of faculty and administrators wielding AI tools, there is a commensurate amount of responsibility to use the AI professionally, equitably, and ethically. AI chat bots may sound human and exhibit a noticeable personality, but they are not persons, they are tools. The writing of AI sounds very intelligent, but that is not because the AI itself is highly intelligent, but rather the AI has assimilated and synthesized the intelligent words of tens of thousands of intelligent writers and generated text that appears equally intelligent. But the AI does not think. It does not reason. It does not replace thinking for our students. AI merely has the extraordinary capacity to simulate the output of a reasonable, thinking person. Yet, AI can also assimilate lies, biased content, hate speech, and harmful language and synthesize and generate similar content without discriminating the good and the true from the bad and the harmful. Current textual generative AIs are trained on a large language models that did not evaluate their source material for truth or bias, or fairness or hatefulness, in gathering as much data as possible for the AI to work with. Volume of material was the operating criteria for building large language models, not truth, justice, equity, and inclusion. At the same time, AI has the capacity to collect and run through personal and biometric data, again without thinking because AI does not think. This is an important part of the use experience of these models, and one that needs to be communicated to users who would turn to the AI for truth and correction on a wide range of deeply important topics.

Download the essay from SSRN at the link.

Duryea on What Generative AI Means for Clinical Legal Education @pelfreyduryea @BU_Law

Danielle Pelfrey Duryea, Boston University School of Law, has published What Does Generative Artificial Intelligence Mean for Clinical Legal Education?

Within the last ten months, OpenAI's ChatGPT and other large language model (LLM) artificial intelligence tools have become widely available, widely celebrated, and widely feared. The legal profession is abuzz with think-pieces on how AI will (or won't) transform the practice of law. Legal academia has begun to digest how these extremely rapidly-evolving tools could change legal education and licensure, including studying practices (e.g., Harrington), law school exams (e.g., Choi, Hickman, Monahan & Schwarcz), legal research and writing (e.g., Romig; Bishop; Baskaran; Smith), podium teaching and academic service (e.g., Pettinato Oltz), and the bar exam (e.g., Williams; Bommarito & Katz). This paper continues this urgent reflection by considering how AI tools change the context for and practice of conventional goals and methods of clinical legal education, including client-centered / collaborative lawyering, from-scratch drafting of legal documents, professional collaboration, etc.

The full text is not available from SSRN. 

A New Blog On AI and Legal Education @theFordon @SeanLovesBooks @JenWondracek @AZSarahMom

Check out AI Law Librarians, this new blog on generative AI and legal education, research, and tech, from some very smart and very entertaining law librarians. 

Lucchi on ChatGPT: A Case Study on Copyright Challenges for Generative AI Systems

Nicola Lucchi, Universitat Pompeu Fabra, Department of Law, has published ChatGPT: A Case Study on Copyright Challenges for Generative AI Systems at 2023 European Journal of Risk Regulation 1. Here is the abstract.

This article focuses on copyright issues pertaining to generative artificial intelligence (AI) systems, with particular emphasis on the ChatGPT case study as a primary exemplar. In order to generate high-quality outcomes, generative artificial intelligence (AI) systems necessitate substantial quantities of training data, which may frequently comprise copyright protected information. This prompts inquiries into the legal principles of fair use, the creation of derivative works, and the lawfulness of data gathering and utilization. The utilization of input data for the purpose of training and enhancing AI models presents significant concerns regarding potential violations of copyright. This paper offers suggestions for safeguarding the interests of copyright holders and competitors, while simultaneously addressing legal challenges and expediting the advancement of artificial intelligence technologies. This study analyzes the ChatGPT platform as a case example to explore the necessary modifications that copyright regulations must undergo to adequately tackle the intricacies of authorship and ownership in the realm of AI-generated creative content. Note: Creative Commons License This work is licensed under a Creative Commons Attribution 4.0 International License.

Download the article from SSRN at the link.

Choi and Schwarcz on AI Assistance in Legal Analysis: An Empirical Study @UofMNLawSchool @Dschwarcz

Jonathan H. Choi and Daniel Schwarcz, both of the University of Minnesota Law School, have published AI Assistance in Legal Analysis: An Empirical Study as Minnesota Legal Studies Research Paper No. 23-22. Here is the abstract.

Can artificial intelligence (AI) augment human legal reasoning? To find out, we designed a novel experiment administering law school exams to students with and without access to GPT-4, the best-performing AI model currently available. We found that assistance from GPT-4 significantly enhanced performance on simple multiple-choice questions but not on complex essay questions. We also found that GPT-4’s impact depended heavily on the student’s starting skill level; students at the bottom of the class saw huge performance gains with AI assistance, while students at the top of the class saw performance declines. This suggests that AI may have an equalizing effect on the legal profession, mitigating inequalities between elite and nonelite lawyers. In addition, we graded exams written by GPT-4 alone to compare it with humans alone and AI-assisted humans. We found that GPT-4’s performance varied substantially depending on prompting methodology. With basic prompts, GPT-4 was a mediocre student, but with optimal prompting it outperformed both the average student and the average student with access to AI. This finding has important implications for the future of work, hinting that it may become advantageous to entirely remove humans from the loop for certain tasks.

Download the article from SSRN at the link.

Volokh on Large Libel Models? Liability for AI Output @unsealiecourt @JournalSpeech

Eugene Volokh, University of California, Los Angeles, School of Law, has published Large Libel Models? Liability for AI Output at 3 Journal of Free Speech Law 489 (2023). Here is the abstract.

“Large Language Model” AI programs routinely invent false and defamatory allegations, complete with invented quotes and invented newspaper articles. Indeed, lawsuits have already been filed over alleged libels created by ChatGPT and Bing. Should such AI programs’ creators be liable for defamation, based on their programs’ output? In this article, I begin by analyzing this question under the current rules of U.S. defamation law. I will tentatively argue that, when the “actual malice” standard applies, the standard might be satisfied if an AI company has received actual notice of particular spurious information being produced by its software but has refused to act. This would in practice require such companies to implement a “notice-and-blocking” system, loosely similar to “notice-and-takedown” systems required under the DMCA as to copyright and trademark infringements. And I will also discuss the possibility of negligence liability, when such liability is authorized under libel law, by analogy to negligent design product liability. To be sure, allowing such liability could yield substantial costs. That is particularly so since it may require lay judges and juries to evaluate complicated technical claims about which designs are feasible. (Such concerns of course mirror the concerns about legal liability as to other products, such as pharmaceuticals or cars, or as to services, such as surgical procedures.) Part II will tentatively discuss some arguments for why the law might be changed, whether by courts, by legislatures, or by administrative agencies. Finally, Part III will offer some similarly tentative thoughts about how this might apply to other claims, such as false light, disclosure of private facts, the right of publicity, or negligence.

Download the article from SSRN at the link.

Iaia on To Be, or Not To Be...Original Under Copyright Law, That Is (One of) the Main Questions Concerning AI-Produced Works @UniLUISS @GrurInt

Vincenzo Iaia, Luiss University, has published To Be, or Not to Be … Original Under Copyright Law, That Is (One of) the Main Questions Concerning AI-Produced Works at 71 GRUR International IX (2022). Here is the abstract.

This paper examines whether different types of AI-produced works, namely AI-assisted, AI-generated, and AI-genassisted works, are able to satisfy the originality requirement, as established within the EU copyright acquis. It will argue that even assuming that the choices of AIs can be deemed free, they could not be considered creative due to the anthropological footprint of copyright law which traditionally covers only human creativity. This implies that copyright protection can be granted only to outputs that display a sufficient level of original human intervention, which is most likely present in AI-assisted works and, depending on the case, in AI-genassisted works, too. The desirability of stretching copyright further to accommodate artificial creativity is also lessened by the risk that it could saturate the creative market due to an overexpansion of monopoly rights. Thus, the copyright solution will be evaluated together with some existing alternative legal tools, such as neighbouring rights, sui generis rights, patents, or even the absence of protection over AI-generated works to preserve a robust public domain.

The full text is not available for download from SSRN.

Asil and Wollmann on Whether Machines Can Commit Crimes Under US Antitrust Laws

Aslihan Asil, Yale University, and Thomas Wollmann, University of Chicago, have published Can Machines Commit Crimes Under US Antitrust Laws? Here is the abstract.

Generative artificial intelligence is being rapidly deployed for corporate tasks including pricing. Suppose one of these machines communicates with the pricing manager of a competing firm, proposes to collude, receives assent, and raises price. Is this a crime under US antitrust laws, and, if so, who is liable? Based on the observed behavior of the most widely adopted large language model, we argue that this conduct is imminent, would satisfy the requirements for agreement and intent under Section 1 of the Sherman Act, and confer criminal liability to both firms as well as the pricing manager of the competing firm.

Download the article from SSRN at the link.