Jamal Awil

SUM.A — SUM.A · Generative AI & Copyright Deep-Dive

SUM pillar · Summer Bridge 2027

A ten-week reading-intensive on the current generative-AI copyright litigation and policy landscape — the USCO AI Report, principal cases, and major scholarly responses. The course that will likely seed the master's thesis topic.

Essential books

• ★ Lemley & Casey — Fair Learning
• ★ Benjamin Sobel — 'AI's Fair Use Crisis'
• ★ Henderson et al. — Foundation Models and Fair Use
• ★ Matthew Sag — 'Copyright Safety for Generative AI'
• ★ Pamela Samuelson — 'Generative AI Meets Copyright'

Supplementary books

• ☆ Annemarie Bridy — Coding Creativity
• ☆ James Grimmelmann — 'Copyright for Literate Robots'
• ☆ Ryan Abbott — The Reasonable Robot
• ☆ Lee et al. — 'Talkin' 'Bout AI Generation'
• ☆ Daniel Gervais — 'The Machine as Author'

Papers (47)

1. Sobel, B. L. W. (2017). Artificial Intelligence's Fair Use Crisis contemporary 2011-2025 Columbia Journal of Law and the Arts, Vol. 41, No. 1, pp. 45–97

A prescient early-era doctrinal analysis predicting that AI training on copyrighted corpora would generate irresolvable fair-use tensions and proposing a compulsory licensing regime as the only structurally coherent solution; cited in virtually every subsequent academic treatment of the training-dat

2. Sag, M. (2009). Copyright and Copy-Reliant Technology established 1990-2010 Northwestern University Law Review, Vol. 103, No. 4, pp. 1607–1682

Establishes the foundational legal category of "copy-reliant technology" — systems that must copy works to function — and argues that intermediate copies made by such systems warrant a distinct, purpose-sensitive fair-use analysis, providing the direct doctrinal precursor to the training-data fair-u

3. Levendowski, A. (2018). How Copyright Law Can Fix Artificial Intelligence's Implicit Bias Problem contemporary 2011-2025 Washington Law Review, Vol. 93, No. 2, pp. 579–630

Argues that copyright law's fair use doctrine, applied to AI training data, inadvertently biases AI systems toward well-documented (and historically white, male, Western) cultural production, proposing that copyright reform can function as an equity tool to force more inclusive training corpus desig

4. Samuelson, P. (1986). Allocating Ownership Rights in Computer-Generated Works canonical pre-1990 University of Pittsburgh Law Review, Vol. 47, No. 4, pp. 1185–1228

The foundational legal-academic treatment of computer authorship, anticipating by four decades the core doctrinal tensions now litigated in Thaler v. Perlmutter and Allen v. Perlmutter; establishes the framework for allocating ownership when machines produce expressive output and remains the essenti

5. Ginsburg, J. C. (2018). People Not Machines: Authorship and What It Means in the Berne Convention contemporary 2011-2025 IIC — International Review of Intellectual Property and Competition Law, Vol. 49, No. 2, pp. 131–135

Argues that the Berne Convention's implicit human authorship requirement forecloses copyright in fully machine-generated works, placing the creative burden on human selection and arrangement; the clearest statement of the international-law case against AI authorship and directly relevant to the USCO

6. Bridy, A. (2012). Coding Creativity: Copyright and the Artificially Intelligent Author contemporary 2011-2025 Stanford Technology Law Review, Vol. 5

Examines whether software-generated creative outputs can qualify for copyright protection under work-for-hire and algorithmic authorship theories, concluding that current doctrine requires human creative expression and anticipating by over a decade the disputes now litigated over generative AI outpu

7. Gervais, D. J. (2020). The Machine as Author contemporary 2011-2025 Iowa Law Review, Vol. 105, No. 5

Proposes a "computational creativity" standard for determining copyrightability of machine-generated works, arguing that courts should assess the degree to which AI systems exhibit originality-like properties rather than applying a strict binary human-author rule; the principal academic counter-argu

8. Ginsburg, J. C. and Budiardjo, L. A. (2019). Authors and Machines contemporary 2011-2025 Berkeley Technology Law Journal, Vol. 34, No. 2

Examines copyright law's treatment of computer-generated works across U.S. and comparative doctrine, arguing that authorship requires a human creative nexus and providing the doctrinal ground-clearing directly applicable to the USCO's AI guidance and to the Thaler and Allen district court decisions.

9. Lee, K., Cooper, A. F., and Grimmelmann, J. (2023). Talkin' 'Bout AI Generation: Copyright and the Generative-AI Supply Chain contemporary 2011-2025 Journal of the Copyright Society of the U.S.A., Vol. 70 (forthcoming; SSRN) (verify)

Maps the generative AI "supply chain" — from data collectors through model trainers to deployers to users — and analyzes how copyright liability should be allocated across each link, arguing that the multi-party structure of modern AI systems requires disaggregated liability analysis that current do

10. Grimmelmann, J. (2016). Copyright for Literate Robots contemporary 2011-2025 Iowa Law Review, Vol. 101, No. 2, pp. 657–681

Reframes the AI-copyright question by arguing that copyright protects communicative expression directed at human readers, not machine processing, offering a doctrinal anchor for analyzing whether LLM training constitutes reading-as-infringement and providing the conceptual framework most frequently

11. Sag, M. (2012). The New Legal Landscape for Text Mining and Machine Learning contemporary 2011-2025 Columbia Journal of Law and the Arts, Vol. 66 (forthcoming; SSRN) (verify) / European Intellectual Property Review (verify)

Surveys the legal environment for text and data mining in both U.S. and EU copyright law immediately before the rise of neural-network-based NLP, identifying the doctrinal gaps that would become central to AI training litigation and arguing that TDM activities are best understood as research-oriente

12. Sobel, B. L. W. (2024). Training Data and the Promise of Copyright Reform (verify) contemporary 2011-2025 Law & Contemporary Problems, Vol. 87 (forthcoming; SSRN) (verify)

Revisits and updates the "AI's Fair Use Crisis" arguments in light of the 2022–2024 wave of litigation, arguing that courts' case-by-case fair use analysis is generating incoherent outcomes and that only a legislative solution — either a TDM exception or a collective licensing regime — can provide t

13. Carlini, N., Ippolito, D., Jagielski, M., Lee, K., Tramer, F., and Zhang, C. (20. Quantifying Memorization Across Neural Language Models contemporary 2011-2025 International Conference on Learning Representations (ICLR 2023)

Empirically demonstrates that large language models memorize and can reproduce verbatim training-data sequences at rates that scale predictably with model size and data repetition, providing the technical grounding for both copyright infringement and privacy arguments about generative AI output and

14. Samuelson, P. (2023). Generative AI Meets Copyright Science, Vol. 381, No. 6654, pp. 158–161

15. Abbott, R. (2020). The Reasonable Robot: Artificial Intelligence and the Law contemporary 2011-2025 Cambridge University Press (book; chapter "AI and Intellectual Property" is the primary academic reference) (verify)

Provides a comparative doctrinal framework examining how tort, patent, and copyright law should adapt to AI-generated outputs, arguing for a "reasonable robot" standard that treats AI as a legally significant agent without granting it personhood — directly applicable to the authorship and training-d

16. Samuelson, P. (2023) (REQUIRED). Generative AI Meets Copyright contemporary 2011-2025 Science, Vol. 381, No. 6654, pp. 158–161

Synthesizes the rapidly evolving copyright questions raised by generative AI — covering training data, AI-output protection, and authorship doctrine — in a major peer-reviewed science venue, making the interdisciplinary stakes of the legal debate accessible to technical, policy, and creative audienc

17. United States Copyright Office (2023–2024). Copyright and Artificial Intelligence: Parts 1, 2, and 3 (USCO AI Report) contemporary 2011-2025 U.S. Copyright Office Policy & International Affairs (official government report; treated as primary source)

The USCO's three-part administrative analysis of AI and copyright covering digital replicas (Part 1), copyrightability of AI-assisted works (Part 2), and AI training data (Part 3); the authoritative administrative statement of current doctrine and the primary policy document framing every live case

18. Thaler v. Perlmutter, No. 22-1564 (D.D.C. Aug. 18, 2023) / aff'd D.C. Cir. 2025. Thaler v. Perlmutter — Academic Commentary: Von Lohmann, F. and others (verify) contemporary 2011-2025 Decisions and academic commentary collected in Journal of the Copyright Society of the U.S.A. (verify)

The foundational AI authorship ruling holding that the Copyright Office properly refused registration for an AI-generated artwork produced without human creative control — establishes the human authorship baseline against which all subsequent AI output copyright claims are measured, including those

19. Allen v. Perlmutter / Allen v. United States Copyright Office (commentary) (2023. Zheng-jie Allen and the Limits of Human Authorship in AI-Assisted Works: Commentary (verify) contemporary 2011-2025 Harvard Journal of Law and Technology Digest / Columbia Journal of Law and the Arts (verify)

Analyzes the Copyright Office's partial registration of Jason Allen's Midjourney-generated "Théâtre D'Opéra Spatial," which turned on whether Allen's iterative prompting constituted sufficient human authorship — establishing the "prompt engineering as selection and arrangement" theory that will gove

20. Zirpoli, C. T. (2023). Generative Artificial Intelligence and Copyright Law contemporary 2011-2025 Congressional Research Service, Report R47877 (2023)

Synthesizes for Congress the state of copyright doctrine as applied to AI training data, AI-assisted works, and AI-generated outputs, serving as the primary U.S. legislative reference framing the USCO's ongoing AI and copyright rulemaking and the clearest entry point for the policy dimension of orig

21. Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991) — c. No "Sweat of the Brow" Copyright for Computer Databases: The Supreme Court Corrects Old Error (commentary) (verify) established 1990-2010 American Journal of Comparative Law, Vol. 41, No. 3 (verify)

Ginsburg's contemporaneous analysis of Feist's originality holding — the constitutional baseline for copyrightability — explains why the "minimum spark of creativity" standard both forecloses most AI-generated output from protection and simultaneously exposes AI companies to claims that their output

22. Tushnet, R. (2024) (verify). The New York Times v. OpenAI and the Future of Journalistic Copyright (verify) contemporary 2011-2025 Columbia Journal of Law and the Arts, Vol. 47 (forthcoming; SSRN) (verify)

Analyzes the NYT v. OpenAI complaint's "regurgitation" theory of infringement, arguing that near-verbatim output reproduction constitutes a viable direct infringement claim that distinguishes this case from broader training-data fair-use disputes and may force doctrinal innovation at the intersectio

23. Rothchild, J. A. (2023) (verify). Andersen v. Stability AI: Artists Challenge the Training of Generative-AI Models on Scraped Images (verify) contemporary 2011-2025 Journal of the Copyright Society of the U.S.A. (forthcoming; SSRN) (verify)

Examines the Andersen v. Stability AI pleadings to assess whether style copying, dataset scraping, and output similarity can independently or jointly support infringement claims, highlighting the doctrinal gaps between visual art and text copyright in AI litigation and the difficulty of proving subs

24. Hughes, J. (2024) (verify). Thomson Reuters v. ROSS Intelligence and the Thinness of Factual Compilations in the AI Era (verify) contemporary 2011-2025 Fordham Intellectual Property, Media and Entertainment Law Journal, Vol. 34 (forthcoming; SSRN) (verify)

Analyzes the Thomson Reuters v. ROSS Intelligence decision holding that AI training on a legal headnote corpus can constitute infringement of thin compilation copyright, exploring what "thin" copyright means when LLMs extract structured knowledge at scale and examining the implications for AI traini

25. Bridy, A. (2024) (verify). AI Training, Fair Use, and the Authors Guild v. OpenAI Litigation (verify) contemporary 2011-2025 George Mason Law Review, Vol. 31 (forthcoming; SSRN) (verify)

Provides the most direct scholarly treatment of Authors Guild v. OpenAI, analyzing whether large-scale ingestion of literary works for LLM training can survive four-factor fair use analysis given the commercial scale and the potential harm to book-licensing markets — directly analogous to the Author

26. Goldman, E. (2024) (verify). Bartz v. Anthropic and the Frontier of AI Output Copyright (verify) contemporary 2011-2025 Santa Clara High Technology Law Journal / Santa Clara Computer and High Technology Law Journal (forthcoming; SSRN) (verify)

Analyzes Bartz v. Anthropic — the first major case directly challenging whether an AI model's conversational outputs can infringe copyright in the conversational corpus it was trained on — examining how Sag's "copyright safety" framework maps onto the specific pleadings and what the case reveals abo

27. Kadrey v. Meta Platforms, Inc., No. 23-cv-03417 (N.D. Cal.) — commentary (2024). Kadrey v. Meta and the Books as Training Data Question (verify) contemporary 2011-2025 Berkeley Technology Law Journal / Journal of the Copyright Society (forthcoming; SSRN) (verify)

Examines the Kadrey v. Meta proceedings in which authors sued Meta over LLaMA training on LibGen-scraped books, analyzing the court's dismissal of certain infringement theories while preserving others and identifying the specific doctrinal thresholds that will control similar book-training claims in

28. Hugenholtz, P. B. (2021). The New Copyright Directive: Text and Data Mining (Articles 3 and 4) (verify) contemporary 2011-2025 Kluwer Copyright Blog / published chapter in Intellectual Property and the Digital Single Market (Wolters Kluwer, 2021) (verify)

Provides the authoritative academic analysis of the DSM Directive's Articles 3 and 4 TDM exceptions — the first binding copyright carve-out for text and data mining in any major jurisdiction — explaining their scope, the opt-out mechanism for commercial mining, and the implications for AI training a

29. Geiger, C., Frosio, G., and Bulayenko, O. (2019). Text and Data Mining in the Proposed Copyright Reform: Making the EU Ready for an Age of Big Data? (verify) contemporary 2011-2025 IIC — International Review of Intellectual Property and Competition Law, Vol. 49 (2018) (verify)

Evaluates the pre-DSM Commission proposals for TDM exceptions against the needs of AI research and commercial machine learning, arguing that the original draft's restrictions (requiring lawful access and research purpose) were too narrow to support industrial AI training and predicting the pressure

30. Bently, L. and Sherman, B. (2019). Article 4 DSM Directive and the Commercial TDM Exception: Scope, Opt-Out, and AI Implications (verify) contemporary 2011-2025 Oxford Research Encyclopedia of Politics / separately as working paper, Cambridge IP (verify)

Provides a close textual analysis of Article 4's commercial TDM carve-out — the provision allowing commercial AI training unless rights holders have "reserved" their rights — explaining how the opt-out mechanism works, what constitutes a valid reservation, and why the resulting two-tier system creat

31. Quintais, J. P. (2020). The New Copyright in the Digital Single Market Directive: A Critical Commentary contemporary 2011-2025 European Intellectual Property Review, Vol. 42, No. 1, pp. 28–41

Provides a comprehensive critical commentary on the full DSM Directive with extended focus on the TDM provisions and Article 17, arguing that the Directive's fragmented approach to AI-relevant copyright creates a patchwork that will require further clarification through CJEU litigation — a predictio

32. Deltorn, J.-M. and Macrez, F. (2018) (verify). Authorship in the Age of Machine Learning and Deep Neural Networks (verify) contemporary 2011-2025 Centre for International Intellectual Property Studies (CEIPI) Research Paper No. 2018-10 (SSRN) (verify)

Maps the doctrinal incompatibility between neural-network creativity and French and EU droit d'auteur traditions, arguing that originality requirements tied to personality expression will systematically exclude deep-learning outputs from protection under civil-law copyright regimes — directly applic

33. Senftleben, M. and Buijtelaar, L. (2020) (verify). Robot Creativity: An Incentive-Based Neighbouring Rights Approach (verify) contemporary 2011-2025 European Intellectual Property Review, Vol. 42, No. 12 (verify)

Proposes that AI-generated outputs be protected under a sui generis "neighbouring right" modeled on EU database rights and sound recording rights rather than under author's rights, providing a European legislative design alternative to U.S.-style authorship analysis and one of the most fully develop

34. Lemley, M. A. (2023) (verify). How Generative AI Turns Copyright Upside Down (verify) contemporary 2011-2025 Stanford Public Law Working Paper (SSRN) (verify)

Updates and extends the "Fair Learning" analysis to account for the commercial generative AI deployment landscape of 2023, arguing that even if training is fair use, output mimicry and market substitution concerns require doctrinal recalibration — Lemley's most direct engagement with the gap between

35. Sag, M. (2024) (verify). Opt-Out Rights and Licensing Frameworks for Generative AI Training Data (verify) contemporary 2011-2025 Law & Contemporary Problems / Houston Law Review supplement (forthcoming; SSRN) (verify)

Evaluates three legislative models for AI training data governance — opt-out exemptions (DSM Article 4 model), voluntary collective licensing, and statutory licensing — against the criteria of transaction cost, creative incentive preservation, and administrative feasibility, concluding that a tiered

36. Cooper, A. F. and Lee, K. (2024) (verify). AI Accountability and the IP System: Toward a Transparency Framework (verify) contemporary 2011-2025 Yale Law Journal Forum / Stanford Technology Law Review (forthcoming; SSRN) (verify)

Argues that copyright doctrine alone is insufficient to govern AI training practices and proposes integrating IP enforcement with AI transparency requirements — mandatory training data disclosure, model cards, and data provenance standards — framing this as an accountability regime that serves both

37. Samuelson, P. (2020). Reconceptualizing Copyright's Merger Doctrine contemporary 2011-2025 Journal of the Copyright Society of the U.S.A., Vol. 63 (verify)

Provides Samuelson's doctrinal analysis of the merger doctrine — the principle that where there are few ways to express an idea, expression merges with idea and is unprotectable — relevant to the reform debate because AI defenders argue that factual and functional content in training corpora is merg

38. Borghi, M. and Karapapa, S. (2013). Copyright and Mass Digitization contemporary 2011-2025 Oxford University Press (book; chapter "Authorizing Computational Uses of Copyright Works" is the primary reference) (verify)

Analyzes the copyright frameworks governing mass digitization and computational analysis of cultural heritage, developing the "authorizing computational use" theory that licenses are not required for non-expressive processing of works — providing a transatlantic doctrinal bridge between Authors Guil

39. Bender, E. M., Gebru, T., McMillan-Major, A., and Shmitchell, S. (2021). On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? contemporary 2011-2025 ACM Conference on Fairness, Accountability, and Transparency (FAccT 2021)

[Bridge: Gen AI Copyright × AI (training data ethics)] Argues that large language models risk encoding and amplifying social biases from training corpora, consuming unsustainable resources, and generating fluent but meaningless text; provides the foundational critical AI framework that contextualize

40. Manovich, L. (2023) (verify). AI Aesthetics (verify) contemporary 2011-2025 Strelka Press / arXiv (verify)

[Bridge: Gen AI Copyright × Communication (AI aesthetics)] Analyzes the aesthetic logic of AI-generated cultural objects, arguing that they represent a qualitatively distinct form of cultural production defined by the "possibility space" of training data rather than individual authorial intent — dir

41. Hao, K. (2023). The Coming War on the Hidden Algorithm (verify) contemporary 2011-2025 MIT Technology Review (long-form journalism; academic citation context: Science, Technology, and Human Values) (verify)

[Bridge: Gen AI Copyright × Communication (Gen AI discourse)] Documents how technology companies frame AI training as a neutral technical process in public discourse, obscuring the legal and ethical significance of training data choices; bridges the copyright debate into media studies by examining h

42. Vats, A. (2020). The Color of Creatorship: Intellectual Property, Race, and the Making of Americans contemporary 2011-2025 Stanford University Press (book; academic review context: Law and Social Inquiry)

[Bridge: Gen AI Copyright × Equity (synthetic blackness / racial IP)] Argues that IP doctrine has historically produced and enforced racial hierarchies by treating white creativity as the default of authorship and innovation — a structural critique directly applicable to whose training data, styles,

43. Birhane, A., Prabhu, V. U., and Kahembwe, E. (2021). Multimodal Datasets: Misogyny, Pornography, and Malignant Stereotypes contemporary 2011-2025 arXiv preprint 2021

[Bridge: Gen AI Copyright × Equity (dataset audit)] Audits LAION-400M and similar large-scale image-text datasets used to train text-to-image models, uncovering systematic misogyny, pornographic content, and racial stereotypes; establishes the dataset audit as a necessary precursor to responsible ge

44. Bommasani, R., Klyman, K., Zhang, D., and Liang, P. (2023). The Foundation Model Transparency Index contemporary 2011-2025 Stanford Center for Research on Foundation Models (CRFM) Technical Report 2023

[Bridge: Gen AI Copyright × Policy (transparency index)] Evaluates ten major foundation models across 100 transparency indicators covering training data, data labor, compute, modalities, and downstream use — producing the most comprehensive empirical dataset on AI transparency and directly supportin

45. Bonadio, E. and McDonagh, L. (2020). Artificial Intelligence as Producer and Consumer of Copyright Works: Evaluating the Consequences of Algorithmic Creativity contemporary 2011-2025 Queen Mary Journal of Intellectual Property, Vol. 10, No. 1, pp. 112–137 (verify)

[Bridge: Gen AI Copyright × International/EU] Provides a comparative UK/EU analysis of AI as both consumer of (training on) and producer of (outputting) copyright works, arguing that the dual role creates doctrinal instability and that the most coherent solution is to treat AI-generated works as own

46. Gervais, D. J. (2019). Exploring the Interfaces Between Big Data and Intellectual Property Law contemporary 2011-2025 Journal of Intellectual Property, Information Technology and E-Commerce Law (JIPITEC), Vol. 10, No. 1

[Bridge: Gen AI Copyright × Policy + AI (big data and IP)] Maps the full set of IP-law interfaces with big data and machine learning — copyright in datasets, trade secret protection for training pipelines, patent eligibility of AI-generated inventions — providing the most comprehensive cross-IP-regi

47. Daly, A., Hagendorff, T., Li, H., Mann, M., Marda, V., Wagner, B., Wang, W., and. Artificial Intelligence, Governance, and Ethics: Global Perspectives contemporary 2011-2025 China AI Governance Project / University of Melbourne Working Paper (SSRN) (verify)

[Bridge: Gen AI Copyright × Policy (global governance)] Surveys AI governance frameworks across the EU, U.S., China, and the Global South, identifying divergent copyright, data, and ethics regimes that will create jurisdictional fragmentation for AI training datasets and outputs; closes the course b

Braided extracurricular reading

Fiction

AI Asimov, I. (1954). The caves of steel. Doubleday.

A noir detective novel pairing a human investigator with a humanoid robot partner, R. Daneel Olivaw, to investigate a murder in a hyperurbanized Earth. The novel explores anti-robot prejudice, labor displacement, and the social-psychological texture of human-machine collaboration—issues now central to HCI, organizational sociology, and AI equity research. Asimov uses the procedural form to ground abstract debates about machine rights and economic transformation in everyday encounters. For scholars interested in narrative framings of human-AI teaming, automation anxiety, and the politics of integration, the book remains a key reference point for thinking through trust, suspicion, and shared agency. AI Le Guin, U. K. (1971). The lathe of heaven. Charles Scribner's Sons.

George Orr's dreams alter reality, and a psychiatrist, Dr. Haber, attempts to wield this power instrumentally to engineer a better world. The novel critiques utopian engineering and technocratic ambition, with explicit resonance for contemporary debates about AI-driven social optimization and behavioral nudging. Le Guin's depiction of recursive consequences, unintended side effects, and the hubris of optimization speaks directly to AI ethics, communication policy, and the politics of welfare prediction. For scholarship on equity, the novel's interrogation of who gets to define the good is enduringly relevant. AI Vinge, V. (2006). Rainbows end. Tor Books.

Set in a near-future San Diego saturated with augmented reality, distributed surveillance, and ubiquitous AI assistance, the novel follows a recovering Alzheimer's patient navigating a transformed world. Vinge dramatizes the social, cognitive, and political consequences of pervasive ambient computing. For scholarship on AI in communication and equity, Rainbows End is unusually prescient about AI-enabled education, generational digital divides, surveillance, and the contest over information infrastructures. AI Bear, G. (1990). Queen of angels. Warner Books.

Set in a near-future Los Angeles, the novel follows multiple storylines—a murder investigation, a deep-space probe encountering a planetary intelligence, and a self-aware AI's emergent consciousness. Bear's careful attention to nanotech-mediated psychiatric therapy and the politics of cognitive modification anticipates contemporary debates over algorithmic mental health interventions. For AI scholarship engaging with consciousness, neurotechnology, and equity, the novel offers an ambitious literary synthesis. AI Banks, I. M. (2010). Surface detail. Orbit.

A Culture novel exploring virtual afterlives, in which simulated Hells house digital subjects in conditions of torture, raising profound ethical questions about the moral status of simulated minds and the politics of computational suffering. For scholarship on AI ethics, posthuman rights, and the equity implications of pervasive simulation, Surface Detail provides an essential literary intervention into debates that current AI ethics has barely begun to address.

Nonfiction

AI Wooldridge, M. (2020). A brief history of artificial intelligence: What it is, where we are, and where we are going. Flatiron Books.

A distinguished AI researcher's chronological history of the field from McCulloch and Pitts to contemporary deep learning. Wooldridge writes with disciplinary authority while remaining accessible to non-specialists, providing essential background on the field's recurrent enthusiasms and disillusionments. For doctoral work engaging with AI policy and communication, Wooldridge's careful periodization of AI winters and springs provides important historical context for evaluating contemporary capability claims. AI Russell, S. (2019). Human compatible: Artificial intelligence and the problem of control. Viking.

Russell's articulation of the "principles of beneficial AI" and his case for inverse reinforcement learning as an alignment strategy. The book is a leading technical-philosophical contribution to AI safety. Essential for doctoral work on AI governance and ethics. AI Benjamin, R. (Ed.). (2019). Captivating technology: Race, carceral technoscience, and liberatory imagination in everyday life. Duke University Press.

An edited volume examining technological systems within carceral and racial-capitalist contexts. Benjamin's editorial framing centers Black scholarly engagement with technology, offering essential resources for equity-AI scholarship. AI O'Neil, C. (2016). Weapons of math destruction: How big data increases inequality and threatens democracy. Crown Publishers.

A foundational popular-critical text on algorithmic harm in credit, education, employment, and policing. O'Neil's framework of WMDs (opaque, scaled, harmful algorithms) has shaped public discourse and policy debate. AI Floridi, L. (2023). The ethics of artificial intelligence: Principles, challenges, and opportunities. Oxford University Press.

Floridi's contemporary synthesis of AI ethics literature. A useful reference for doctoral work mapping the AI ethics landscape.

---

SUM.B — SUM.B · Technical Foundations of AI for Policy Thinkers

SUM pillar · Summer Bridge 2027

A working-through of the technical content a serious policy writer must hold — gradient descent, backpropagation, attention, scaling laws, alignment, evaluation, agentic systems — without becoming an ML engineer. MOOCs paired with hands-on conceptual exercises.

Essential books

• ★ Aurélien Géron — Hands-On Machine Learning
• ★ Goodfellow, Bengio & Courville — Deep Learning
• ★ Bommasani et al. — Foundation Models* (CRFM)
• ★ Bender et al. — Stochastic Parrots
• ★ Stephen Wolfram — What Is ChatGPT Doing?

Supplementary books

• ☆ Kevin Murphy — Probabilistic ML
• ☆ Chip Huyen — Designing ML Systems
• ☆ Martin Kleppmann — Designing Data-Intensive Apps
• ☆ Hastie, Tibshirani & Friedman — ESL
• ☆ Christopher Bishop — Pattern Recognition and ML

Papers (50)

1. Domingos, Pedro. "A Few Useful Things to Know About Machine Learning." *Communic. **

2. LeCun, Yann, Yoshua Bengio, and Geoffrey Hinton. "Deep Learning." Nature 521 (. **

3. ?. Mitchell, Tom M. "The Discipline of Machine Learning." Carnegie Mellon University Technical Report CMU-ML-06-108, 2006.

4. ?. Hastie, Trevor, Robert Tibshirani, and Jerome Friedman. "Overview of Supervised Learning." Chapter 2 in *The Elements of Statistical Learnin

5. Jordan, Michael I., and Tom M. Mitchell. "Machine Learning: Trends, Perspectives. **

6. Sculley, D., Gary Holt, Daniel Golovin, Eugene Davydov, Todd Phillips, Dietmar E. **

7. ?. Rahimi, Ali, and Benjamin Recht. "Reflections on Random Kitchen Sinks." NIPS Test-of-Time Award Talk, 2017. (Transcript / slide deck.)

8. Krizhevsky, Alex, Ilya Sutskever, and Geoffrey E. Hinton. "ImageNet Classificati. **

9. ?. He, Kaiming, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. "Deep Residual Learning for Image Recognition." *Proceedings of the IEEE Conference

10. Goodfellow, Ian, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, S. **

11. Hochreiter, Sepp, and Jürgen Schmidhuber. "Long Short-Term Memory." *Neural Comp. **

12. ?. Olah, Christopher. "Neural Networks, Manifolds, and Topology." Distill.pub, 2014.

13. ?. Olah, Christopher, Arvind Satyanarayan, Ian Johnson, Shan Carter, Ludwig Schubert, Katherine Ye, and Alexander Mordvintsev. "The Building Bl

14. Silver, David, Aja Huang, Chris J. Maddison, Arthur Guez, Laurent Sifre, George. **

15. Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan. **

16. ?. Devlin, Jacob, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. "BERT: Pre-training of Deep Bidirectional Transformers for Language Under

17. ?. Radford, Alec, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. "Language Models Are Unsupervised Multitask Learners." Op

18. Raffel, Colin, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael. **

19. Brown, Tom, et al. (OpenAI). "Language Models Are Few-Shot Learners." *Advances. **

20. Wei, Jason, Yi Tay, Rishi Bommasani, Colin Raffel, Barret Zoph, Sebastian Borgea. **

21. Schuster, Mike, and Kaisuke Paliwal. "Bidirectional Recurrent Neural Networks.". **

22. ?. Bommasani, Rishi, Drew A. Hudson, Ehsan Adeli, Russ Altman, Simran Arora, Sydney von Arx, Michael S. Bernstein, et al. "On the Opportunities

23. ?. Kaplan, Jared, Sam McCandlish, Tom Henighan, Tom B. Brown, Benjamin Chess, Rewon Child, Scott Gray, Alec Radford, Jeffrey Wu, and Dario Amod

24. ?. Hoffmann, Jordan, Sebastian Borgeaud, Arthur Mensch, Elena Buchatskaya, Trevor Cai, Eliza Rutherford, Diego de Las Casas, et al. "Training C

25. Ouyang, Long, Jeff Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishk. **

26. ?. Touvron, Hugo, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, et al. "LLaMA: Open

27. ?. Bubeck, Sébastien, Varun Chandrasekaran, Ronen Eldan, Johannes Gehrke, Eric Horvitz, Ece Kamar, Peter Lee, et al. "Sparks of Artificial Gene

28. ?. Anthropic. "Claude's Model Card." Anthropic, 2023 (updated). Online document.

29. Gebru, Timnit, Jamie Morgenstern, Briana Vecchione, Jennifer Wortman Vaughan, Ha. **

30. ?. Sambasivan, Nithya, Shivani Kapania, Hannah Highfill, Diana Akrong, Praveen Paritosh, and Lora M. Aroyo. "'Everyone Wants to Do the Model Wo

31. ?. Birhane, Abeba, Vinay Uday Prabhu, and Emmanuel Kahembwe. "Multimodal Datasets: Misogyny, Pornography, and Malicious Stereotypes." arXiv:211

32. Paullada, Amandalynne, Inioluwa Deborah Raji, Emily M. Bender, Emily Denton, and. **

33. ?. Dodge, Jesse, Maarten Sap, Ana Marasović, William Agnew, Gabriel Ilharco, Dirk Groeneveld, Margaret Mitchell, and Matt Gardner. "Documenting

34. ?. Luccioni, Alexandra Sasha, and Joseph D. Viviano. "What's in the Box? An Analysis of Undesirable Content in the Common Crawl Corpus." arXiv:

35. Bender, Emily M., and Batya Friedman. "Data Statements for Natural Language Proc. **

36. Liang, Percy, Rishi Bommasani, Tony Lee, Dimitris Tsipras, Dilara Soylu, Michihi. **

37. ?. Bowman, Samuel, and George Dahl. "What Will It Take to Fix Benchmarking in Natural Language Understanding?" Proceedings of NAACL 2021, pp.

38. ?. Birhane, Abeba, Pratyusha Kalluri, Dallas Card, William Agnew, Ravit Dotan, and Michelle Bao. "The Values Encoded in Machine Learning Resear

39. Lipton, Zachary C., and Jacob Steinhardt. "Troubling Trends in Machine Learning. **

40. ?. Raji, Inioluwa Deborah, Emily M. Bender, Amandalynne Paullada, Emily Denton, and Alex Hanna. "AI and the Everything in the Whole Wide World

41. ?. Anthropic. "Responsible Scaling Policy." Anthropic, 2023. Online document.

42. ?. Hutchinson, Ben, Vinodkumar Prabhakaran, Emily Denton, Kellie Webster, Yu Zhong, and Stephen Denuyl. "Evaluation Gaps in Machine Learning Pr

43. ?. Bender, Emily M., Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. "On the Dangers of Stochastic Parrots: Can Language Mode

44. ?. Marcus, Gary. "Deep Learning Is Hitting a Wall." Nautilus, March 2022.

45. ?. Crawford, Kate, and Vladan Joler. "Anatomy of an AI System: The Amazon Echo as an Anatomical Map of Human Labor, Data, and Planetary Resourc

46. ?. Bender, Emily M., and Alexander Koller. "Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data." *Proceedings of ACL

47. ?. Buolamwini, Joy, and Timnit Gebru. "Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification." *Proceedings of

48. Birhane, Abeba. "Algorithmic Injustice: A Relational Ethics Approach." *Patterns. **

49. ?. Henderson, Peter, Xuechen Li, Dan Jurafsky, Tatsunori Hashimoto, Mark A. Lemley, and Percy Liang. "Foundation Models and Fair Use." arXiv:23

50. ?. National Institute of Standards and Technology. "Artificial Intelligence Risk Management Framework (AI RMF 1.0)." NIST AI 100-1. Gaithersbur

Braided extracurricular reading

Fiction

AI Asimov, I. (1957). The naked sun. Doubleday.

In this sequel, Elijah Baley investigates a murder on Solaria, a thinly populated world where citizens live in isolation, mediated entirely by robotic servants and telepresence. The novel anticipates contemporary anxieties about remote work, social atomization, and the affective consequences of mediated interaction—now central to communication studies of telework and AI companionship. Asimov's depiction of a society wholly dependent on robotic infrastructure dramatizes the political economy of automation and the cultural reconfiguration of intimacy. The text remains pertinent for analyzing how AI infrastructures restructure proximity, sociality, and care, with implications for equity, labor, and the politics of presence. AI Le Guin, U. K. (1974). The dispossessed. Harper & Row.

A physicist from an anarchist moon travels to its capitalist parent planet to share a theory that could enable instantaneous communication. The novel examines the politics of knowledge, ownership, and dissemination—directly relevant to IP scholarship and to debates about open science and the commons of AI models and datasets. Le Guin's juxtaposition of property regimes illuminates how knowledge institutions shape who benefits from technical breakthroughs. For doctoral work at the AI–IP–equity nexus, The Dispossessed is canonical for thinking about commons, intellectual property, and the social architecture of scientific communication. AI Stephenson, N. (1992). Snow crash. Bantam Books.

The novel's depiction of the Metaverse, linguistic viruses, and corporatized governance frames the cyberpunk imagination of the 1990s. Stephenson's interrogation of language as code—and code as biopolitical infrastructure—remains generative for thinking about large language models, prompt injection, and adversarial inputs. For AI scholarship, the novel offers a literary apparatus for theorizing the entanglement of language, computation, and power, with direct implications for communication studies, IP, and equity. AI Egan, G. (1994). Permutation city. Millennium.

A novel that takes seriously the philosophical implications of computational substrate independence, exploring uploaded minds, simulated cosmologies, and the dust theory of consciousness. Egan's careful philosophical work makes Permutation City essential reading for those engaging with the metaphysics of machine consciousness, the simulation argument, and the ethics of digital persons. For doctoral research on AI personhood, IP rights in synthetic minds, and the legal status of computational subjects, the novel offers an unusually rigorous narrative thought experiment. AI Mieville, C. (2011). Embassytown. Macmillan.

A linguistic science fiction novel in which human ambassadors learn to speak a language that requires two simultaneous voices to address an alien species whose speech cannot lie. Mieville stages questions about the politics of translation, semantic incommensurability, and the materiality of communication—directly relevant to contemporary debates about machine translation, large language models, and AI-mediated cross-cultural communication. For communication scholarship at the AI intersection, Embassytown is generative for theorizing linguistic mediation across radical difference.

Nonfiction

AI Marcus, G., & Davis, E. (2019). Rebooting AI: Building artificial intelligence we can trust. Pantheon Books.

A sustained critique of contemporary deep learning paradigms by two cognitive scientists, who argue that genuine progress requires the integration of symbolic and connectionist approaches. Marcus and Davis identify common-sense reasoning, causality, and compositional generalization as persistent weaknesses in contemporary systems. For policy-relevant AI scholarship, the book is essential for grounding capability claims in a clear-eyed view of contemporary system limitations. AI Bostrom, N. (2014). Superintelligence: Paths, dangers, strategies. Oxford University Press.

The canonical statement of the existential-risk argument concerning AI. Bostrom's careful philosophical analysis of intelligence explosion, instrumental convergence, and value loading has shaped subsequent AI safety research. Essential reading at the AI–policy–philosophy intersection. AI Noble, S. U. (2018). Algorithms of oppression: How search engines reinforce racism. New York University Press.

Noble's foundational analysis of how search engine algorithms reproduce racial domination, with sustained attention to Black women's representation. Indispensable for AI-communication-equity scholarship. AI Broussard, M. (2018). Artificial unintelligence: How computers misunderstand the world. MIT Press.

Broussard's analysis of "technochauvinism" and the persistent limitations of AI systems. Essential for AI-equity-communication scholarship. AI Birhane, A. (2022). Algorithmic injustice: A relational ethics approach. (Doctoral dissertation work and related papers published in venues including Patterns and Big Data & Society).

While not a single book, Birhane's published corpus including her doctoral dissertation provides foundational scholarship on algorithmic colonialism, relational ethics in AI, and the politics of data. For doctoral work engaging with critical AI studies from a Global South perspective, Birhane is essential. (Doctoral candidates should also engage Birhane's articles, especially "Algorithmic Colonization of Africa.")