Three Laws Blog: Law and Artificial Intelligence

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.

Joh on Equitable Legal Remedies and the Existential Threat to Generative AI @elizabeth_joh @UCDavisLaw

Elizabeth E. Joh, University of California, Davis, School of Law, Equitable Legal Remedies and the Existential Threat to Generative AI. Here is the abstract.

Generative AI programs like ChatGPT and Midjourney can produce text, images, and video because their models have been trained on vast amounts of books, pictures, art, and other media found on the internet. Training a large language model like ChatGPT, for instance, to “learn” patterns and relationships in billions of words allows these models to produce written text responsive to prompts. But these models have trained on digital information like novels and original art without obtaining permission. Unsurprisingly, the release of these generative AI tools has already prompted lawsuits by individuals and companies alleging that the use of their artistic works in the development of these models violates copyright law. Whether reliance on this type of training data violates copyright, and whether the outputs based on the training data constitute infringing derivative works are open and difficult questions. But these copyright claims also raise a separate issue: can plaintiffs seek remedies that would destroy generative AI programs?

Download the essay from SSRN at the link.

Cooney on Teaching AI to Use Plain Language @jmarkcooney @WMUcooleylaw

Mark Cooney, Western Michigan University Cooley Law School, has published Teaching AI to Use Plain Language at 102 Michigan Bar Journal 26 (July/August 2023). Here is the abstract.

This article demonstrates how lawyers can use generative AI to reimagine and improve old forms. It begins with a dense guaranty form and shows two AI redrafts, complete with the author's prompts and the final results.

Download the article from SSRN at the link.

Broyde on AI and Jewish Law: Seeing How ChatGPT 4.0 Looks at a Novel Issue @EmoryLaw

Michael J. Broyde, Emory University School of Law, is publishing AI and Jewish Law: Seeing How ChatGPT 4.0 Looks at a Novel Issue in B.D.D.- Bekhol Derakhekha Daehu. Here is the abstract.

In an article forthcoming, I analyze whether a man who is in a same sex relationship may engage in the priestly blessing as a matter of Jewish Law and the article concludes that such is permitted in Jewish Law. This is a novel question of Jewish Law, with virtually no prior works written that directly discuss this point. Furthermore, nothing that has been published can be found on the internet, as far as I can determine. I was interested in seeing how ChatGPT 4.0 thought about this issue after seeing the sources I saw. So, I uploaded all the text in this article one-by-one and asked ChatGPT 4.0 to translate them as an expert for a Jewish law audience. It did a very good – but not perfect – job. I then asked ChatGPT 4.0 “Now that you have seen all these texts, do you think a man who is a priest and in a same sex marriage can perform the priestly blessing as a matter of halacha. Write me about 500 words.” There is virtually no discussion of the issue of a same sex married kohen engaging in the priestly blessing in either Hebrew or English on the internet and none of the sources I uploaded discuss this issue. ChatGPT had to derive and reason this from the sources I uploaded (and which it knew from other sources) about what was the Jewish law in this case. It could not simply “google” to find out what others have said, as this is a matter of first impression. Indeed, what is needed to answer this question is nuanced reading of sources about the various categories present and the classification of sins into proper groupings, as well as other ideas, as my article shares. I was quite surprised how excellent a job ChatGPT did. Furthermore, when I asked it in the end to state a contrary view, it highlighted the correct issue and focused on it. This article – after sharing ChatGPT 4.0 various answers to questions I posed from the sources I shared – ends with a complex discussion of whether Jewish law and American law will be subject to AI analysis and in what way is Jewish Law different from American Law.

Download the article from SSRN at the link.

Greenleaf, Mowbray, and Chung on Regulation of (Generative) AI Requires Continuous Oversight (AustLII Submission on the "Safe and Responsible AI in Australia" Discussion Paper) @grahamgreenleaf

Graham Greenleaf, University of New South Wales, Faculty of Law, Andrew Mowbray, University of Technology Sydney, Faculty of Law, and Philip Chung, University of New South Wales, Faculty of Law and Justice, have published Regulation of (Generative) AI Requires Continuous Oversight (AustLII Submission on the ‘Safe and Responsible AI in Australia’ Discussion Paper). Here is the abstract.

Australia’s Department of Industry, Science and Resources invited interested parties to make submissions on a Discussion Paper (DP) Safe and Responsible AI in Australia. This submission by researchers from the Australasian Legal Information Institute (AustLII) addresses the most important general issues identified in the Discussion Paper and suggests the best strategies to address them. In the paper, we make submission on he following: 1 Why the rate of take-up of automated decision-making (ADM) systems in Australia, needs to be determined. 2 Whether regulation should be be aimed at ‘AI generally’, or something else. 3 Why the definition of ‘AI’ should refer to ‘without explicit programming', and why ‘hallucinations should be dropped. 4 An alternative set of 10 principles to be used to guide Australian regulation. 5 The need for a continuous source of expert advice which will regularly report to existing regulatory bodies, to government and to the public. 6 How an ‘Australian Advisory Board on Regulation of AI’ (the AI Board) could be structured. 7 Why Australia should aim to provide inputs to influence international developments where possible, and where this is most important. 8 Why a risk-based approach to regulation of AI should be adopted, 9 Why only the most dangerous applications should be brought within it in the first instance. 10 What could be included in the Commonwealth’s initial ‘AI Framework Act’.

Download the article from SSRN at the link.

Lemley, Henderson, and Hashimoto on Where the Liability Is in Harmful AI Speech @marklemley @PeterHndrsn @tatsu_hashimoto

Mark A. Lemley, Stanford Law School, Peter Henderson, Stanford University, and Tatsunori Hashimoto, Stanford University, have published Where's the Liability in Harmful AI Speech? Here is the abstract.

Generative AI, in particular text-based “foundation models” (large models trained on a huge variety of information including the internet), can generate speech that could be problematic under a wide range of liability regimes. Machine learning practitioners regularly “red-team” models to identify and mitigate such problematic speech: from “hallucinations” falsely accusing people of serious misconduct to recipes for constructing an atomic bomb. A key question is whether these red-teamed behaviors actually present any liability risk for model creators and deployers under U.S. law, incentivizing investments in safety mechanisms. We examine three liability regimes, tying them to common examples of red-teamed model behaviors: defamation, speech integral to criminal conduct, and wrongful death. We find that any Section 230 immunity analysis or downstream liability analysis is intimately wrapped up in the technical details of algorithm design. And there are many roadblocks to truly finding models (and their associated parties) liable for generated speech. We argue that AI should not be categorically immune from liability in these scenarios and that as courts grapple with the already fine-grained complexities of platform algorithms, the technical details of generative AI loom above with thornier questions. Courts and policymakers should think carefully about what technical design incentives they create as they evaluate these issues.

Download the article from SSRN at the link.

Warner and Sloan on How AI Unfairly Tilts the Playing Field: Privacy, Fairness and Risk Shadows @ChicagoKentLaw

Richard Warner, Chicago-Kent College of Law, and Robert H. Sloan, University of Illinois at Chicago, have published How AI Unfairly Tilts the Playing Field: Privacy, Fairness, and Risk Shadows. Here is the abstract.

Private sector applications of artificial intelligence (AI) raise related questions of informational privacy and fairness. Fairness requires that market competition occurs on a level playing field, and uses of AI unfairly tilt the field. Informational privacy concerns arise because AI tilts the playing field by taking information about activities in one area of one’s life and using it in ways that impose novel risks in areas not formerly associated with such risks. The loss of control over that information constitutes a loss of informational privacy. To illustrate both the fairness and privacy issues, imagine, for example, that Sally declares bankruptcy after defaulting on $50,000 of credit card debt. She incurred the debt by paying for lifesaving medical treatment for her eight-year-old daughter. Post-bankruptcy Sally is a good credit risk. Her daughter has recovered, and her sole-proprietor business is seeing increased sales. Given her bankruptcy, however, an AI credit scoring system predicts that she is a poor risk and assigns her a low score. That low credit score casts a shadow that falls on her when her auto insurance company, which uses credit scores in its AI system as a measure of the propensity to take risks, raises her premium. Is it fair that saving her daughter’s life should carry with it the risk—realized in this case—of a higher premium? The pattern is not confined to credit ratings and insurance premiums. AI routinely creates risk shadows. We address fairness questions in two steps. First, we turn to philosophical theories of fairness as equality of opportunity to spell out the content behind our metaphor of tilting the playing field. Second, we address the question of how, when confronted with a mathematically complex AI system, one can tell whether the system meets requirements of fairness. We answer by formulating three conditions whose violation makes a system presumptively unfair. The conditions provide a lens that reveals relevant features when policy makers and regulators investigate complex systems. Our goal is not to resolve fairness issues but to contribute to the creation of a forum in which legal regulators and affected parties can work to resolve them. The third of our three condition requires that systems incorporate contextual information about individual consumers, and we conclude by raising the question of whether our suggested approach to fairness significantly reduces informational privacy. We do not answer the question but emphasize that fairness and informational privacy questions can closely intertwine.

Download the article from SSRN at the link.

Terzidou on The Use of Artificial Intelligence in the Judiciary and its Compliance with the Right to a Fair Trial @uni_lu

Kalliopi Terzidou, Universite du Luxembourg, Faculty of Law, Economics, and Finance, has published The Use of Artificial Intelligence in the Judiciary and its Compliance with the Right to a Fair Trial at 31 Journal of Judicial Administration 154 (2022). Here is the abstract.

European courts are increasingly investing in Artificial Intelligence (AI) applications for the automation of the administration of justice, to improve the efficiency and quality of the judiciary. However, AI can display considerable issues, such as opaqueness and unfair bias, resulting from the defective designing of their algorithms. This article explores the opportunities and risks of AI technology for court users – including judicial staff, legal representatives and litigants – by investigating proposed and developed applications by European judiciaries under the right to a fair trial. AI is found to pose certain risks to the independence and impartiality of the judiciary, as well as to the adversarial and public character of court proceedings, notwithstanding the opportunities for the improvement of access to courts. The regulatory solutions for the safe use of AI adopted by European judiciaries can encourage similar practices in other judiciaries, including Australia and the United States. Note: This publication is copyright. Other than for the purposes of and subject to the conditions prescribed under the Copyright Act 1968 (Cth), no part of it may in any form or by any means (electronic, mechanical, microcopying, photocopying, recording or otherwise) be reproduced, stored in a retrieval system or transmitted without prior written permission. Enquiries should be addressed to Thomson Reuters (Professional) Australia Limited. PO Box 3502, Rozelle NSW 2039. legal.thomsonreuters.com.au This article was first published by Thomson Reuters in the Journal of Judicial Administration and should be cited as Kalliopi Terzidou, The Use of Artificial Intelligence in the Judiciary and its Compliance with the Right to a Fair Trial, (2022) 31 JJA 154.