The DOI matches its Crossref record.
How Generative AI Turns Copyright Upside Down
Abstract
While courts are litigating many copyright issues involving generative AI, from who owns AI-generated works to the fair use of training to infringement by AI outputs, the most fundamental changes generative AI will bring to copyright law don't fit in any of those categories. The new model of creativity generative AI brings puts considerable strain on copyright’s two most fundamental legal doctrines: the idea-expression dichotomy and the substantial similarity test for infringement. Increasingly creativity will be lodged in asking the right questions, not in creating the answers. Asking questions may sometimes be creative, but the AI does the bulk of the work that copyright traditionally exists to reward, and that work will not be protected. That inverts what copyright law now prizes. And because asking the questions will be the basis for copyright ability, similarity of expression in the answers will no longer be of much use in proving the fact of copying of the questions. That means we may need to throw out our test for infringement, or at least apply it in fundamentally different ways.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Figures

![Tan KW, Jailbreaking AI Chatbots is Tech's New Pastime , I3INVESTOR (Apr. 9, 2023, 12:29 PM), https://klse.i3investor.com/web/blog/detail/kianweiaritcles/2023-04-09-story-h-274132588-Jailbreaking_AI_chatbots_is_tech_s_new_pastime [https://perma.cc/XFN7-CSHQ].](/paper-figures/b26dd98c7aff453dbcea190f6544b8e7/how-generative-ai-turns-copyright-upside-down-figure-2.png)




![48 See Maria Diaz, How to Use ChatGPT , ZDNet, June 30, 2023, https://www.zdnet.com/article/how-to-use-chatgpt/ [https://perma.cc/M529-ES7Q] ('Most of the time, when different people ask chatGPT the same question, they will get the same answer. There might be a few variations in words, but the response will be almost identical.').](/paper-figures/0394d0c9065d4d73a86f2c07404993e6/how-generative-ai-turns-copyright-upside-down-figure-7.png)
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“In the context of GenAI, the test of”
From Redefining boundaries in innovation and knowledge domains: Investigating the impact of generative artificial intelligence on copyright and intellectual property rights (2024) · cites this paper as Lemley, 2023
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
No references could be resolved yet.
Research works printed but not linked: 17
- 2023, In the case of works containing AI-generated material, the Office will consider whether the AI contributions are the result of 'mechanical reproduction' or instead of an author's 'own original mental conception, to which [the author] gave visible form
- Regina See, 2006, ] (noting that Chihuly, Warhol, Koons and others rely on others to produce their works). I am indebted to Matt, SEATTLE PI
- at 1680 ("[R]ather than lowering the cost of appropriability for works that are expensive to create, AIs make initial creation itself inexpensive
- 2023, Lee also notes that authorship may arise in the selection of elements of a work, though presumably he would not extend that to, say, selecting one of the four images Midjourney or Dall-E offers in response to a prompt. Id. at 60-62. 37 And quite good, as it happens. See Tim Marcin, This Fake Song from Drake and The Weeknd is AI-Generated and Scary Good, MASHABLE, 75
- 1976, it disappeared almost immediately as a result of (somewhat dubious) copyright takedown notices. You can, however, find the song without their voices-something that is entirely generated by AI. Spectre, Heart On My The open source world is an obvious counterexample, but the proprietary software industry struggled for years with how to cope with it, William Henry Gates III, An Open Letter To Hobbyists
- Amy Harmon, 1998, Internal Memo Shows Microsoft Executives' Concern Over Free Software, N.Y. TIMES
- Mitch Wagner, 2021, explaining that an open-source version of Java, OpenJDK, only became available in 2007 because "in the 1990s the tech industry wasn't as accepting of open source as it is today, Is Java SE Open Source Software? The Short Answer Is 'Yes
- E G, 2023, OpenAI] hereby assign[s] to you all [its] right, title and interest, if any, in and to Output, GitHub Copilot Product Specific Terms
- 2023, See generally Andy Warhol Found. for the Visual Arts, Inc. v. Goldsmith, 598 U.S
- F Supp, 2005, 72 To be clear, though, we cannot expect-and would not want-to have only machine creativity. Among other problems, generative AI works well only if it is trained on human-generated content, The Curse of Recursion: Training on Generated Data Makes Models Forget, 453
- Burk, supra note 2, at 1693 (arguing that AI will drive social concern over the value of authenticity
- Zahr Said, 2023, Fables of Scarcity in IP, FRONTIERS IN RSCH. METRICS & ANALYTICS, 1
- Cf, supra note 2, at 1708-09 (noting this fact but also noting the poor fit between trademark law and authenticity in a social sense)
- Lydia M Hopper, Although animals have not created the diversity and complexity of material technologies that humans have . . . they nonetheless innovate. Innovations by animals have been reported in a range of realms including tool construction and use . . . the eating and processing of novel foods . . . and the invention of novel gestures to communicate meaning and maintain social bonds
- 2012, TECH, Annemarie Bridy, Coding Creativity: Copyright and the Artificially Intelligent Author, 1
- 2019, Artificial Intelligence (AI), specifically GPT-4, was found to match the top 1% of human thinkers on a standard creativity test. The AI application ChatGPT, developed using GPT-4, excelled in fluency and originality in the Torrance Tests of Creative Thinking, a widely recognized tool for assessing creativity, Sean Dorrance Kelly, A Philosopher Argues That an AI Can't Be an Artist, MIT TECH. REV
- Keith Kirkpatrick, 2023, Despite the ability of AI to produce creative outputs based on the attributes of existing works, the process is not the same as a human's creativity, which comes from a combination of real-world experience, emotion, and inspiration, Can AI Demonstrate Creativity?, 66 COMMC'NS OF THE ACM, 21
Cited by
Al-Busaidi, A. S., Raman, R., Hughes, L., Albashrawi, M. A., Malik, T., Dwivedi, Y. K., Al- Alawi, T., AlRizeiqi, M., Davies, G., Fenwick, M., et al. (2024) · Journal of Innovation & Knowledge · doi:10.1016/j.jik.2024.100630 · Cited by 95
Metadata
- Authors
- Lemley, M. A.
- ORCID
- Lemley, Mark A.
- Publisher
- Elsevier BV
- Published
- 2023
- Cited by
- 17
- References
- 0
- Type
- Preprint
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review