Intro class on “The Impact of Generative AI on Critical Thinking”

Hao-Ping (Hank) Lee, Advait Sarka, Lev Tankelevitch, Ian Drosos, Sean Rintel, Richard Banks, and Nicholas Wilson, The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers, available at https://www.microsoft.com/en-us/research/publication/the-impact-of-generative-ai-on-critical-thinking-self-reported-reductions-in-cognitive-effort-and-confidence-effects-from-a-survey-of-knowledge-workers.

This is a synthetic introductory class (no expertise assumed!) generated using enTalkenator (selecting Claude 3.7 Sonnet).

Abstract: “The rise of Generative AI (GenAI) in knowledge workflows raises questions about its impact on critical thinking skills and practices. We survey 319 knowledge workers to investigate 1) when and how they perceive the enaction of critical thinking when using GenAI, and 2) when and why GenAI affects their effort to do so. Participants shared 936 first-hand examples of using GenAI in work tasks. Quantitatively, when considering both task- and user-specific factors, a user’s task-specific self-confidence and confidence in GenAI are predictive of whether critical thinking is enacted and the effort of doing so in GenAI-assisted tasks. Specifically, higher confidence in GenAI is associated with less critical thinking, while higher self-confidence is associated with more critical thinking. Qualitatively, GenAI shifts the nature of critical thinking toward information verification, response integration, and task stewardship. Our insights reveal new design challenges and opportunities for developing GenAI tools for knowledge work.”

Workshop on Hannah Walser’s “Interpretive Facts: Textualism, Empiricism, and the Law-Fact Divide”

Hannah Walser, Interpretive Facts: Textualism, Empiricism, and the Law-Fact Divide.

Solum’s Download of the Week for May 17, 2025. Available on SSRN.

This is a synthetic academic workshop generated using enTalkenator (a variation of the Workshop template, using Claude 3.7 Sonnet).

Abstract: “This Article offers the first systematic examination of where empirical linguistic evidence should fit in the federal court system's division of labor based on law and fact. I begin by identifying "interpretive facts": generalized empirical evidence about how a particular word or phrase has been used by a population of speakers, understood by a population of listeners, or both. Thanks to technological advances in corpus linguistics permitting the quantitative analysis of large numbers of texts, as well as methodological shifts that frame ordinary meaning as an empirical question, such facts are increasingly prominent in statutory and constitutional interpretation cases. I explain the complications that interpretive facts create for two key aspects of the law-fact division of labor: the default expectation that questions of fact be submitted to a jury, and the more deferential standard of appellate review that applies to a trial court's findings of fact. If courts continue to rely on interpretive facts to interpret statutes and constitutional provisions, I argue that they should avoid procedural incoherence by drawing on techniques developed in two other contexts: contract and patent law. Linguists offering interpretive fact evidence should present their arguments to a jury or a judge acting as fact-finder, as in contract cases; findings of interpretive fact should be reviewed for clear error by appellate courts, as in patent claim construction cases. In closing, I situate interpretive facts in the context of scholarly conversations about the increasingly dispositive role of historical facts in the Supreme Court's jurisprudence.”

Workshop on E. Garrett West, “Constitutional Private Law”

E. Garrett West, Constitutional Private Law.

Solum’s Download of the Week for May 10, 2025. Available on SSRN.

This is a synthetic academic workshop generated using enTalkenator (a variation of the Workshop template, using Google Gemini 2.5 Pro).

Abstract: “Constitutional private law is like ordinary private law.  It imposes relational obligations on government officials, like duties to use only reasonable force against a person or not to fire an employee for discriminatory reasons, that are analogous to the rules of tort, property, or contract.  Constitutional public law, by contrast, controls when and how governments validly change non-constitutional legal rules.  When a government violates constitutional public law, as when Congress exceeds its Article I powers, the government fails to change non-constitutional legal rules but does not breach a relational duty to anyone.  That difference should reframe both forms of constitutional law.  It explains why constitutional remedies might either under-or over-protect the interests safeguarded by constitutional public law. And it shows that the best defense of constitutional private law draws on a vision of ordinary private law as a system for redressing wrongs, and that constitutional private law should be less concerned about constitutional theory and more concerned about articulating basic norms of interpersonal justice for those with official authority.”

Interdisciplinary Workshop on “Trophic level influences larval Shortbelly Rockfish development”

Kwan et al., Trophic level influences larval Shortbelly Rockfish development, available at https://academic.oup.com/mcf/article/16/6/e10319/7959870.

This is a synthetic academic workshop generated using enTalkenator (using an AI-generated interdisciplinary workshop template and authored by Google Gemini 2.5 Pro).

Abstract: “Objective: Early life success of fishes is considered one of the most important drivers of recruitment to adult populations, and elucidating the governing mechanisms is important for management efforts. Many hypotheses over the past century have been proposed to explain recruitment fluctuation, with the recently postulated Trophic Efficiency in Early Life (TEEL) hypothesis arguing that a shorter food chain length equals greater energy transfer efficiency from primary producers to larval fishes, thereby reducing early-­ life mortality and ultimately leading to stronger recruitment. Under TEEL it would then be assumed that feeding low in the food chain would improve growth and body condition, as these are often shown to be associated with increased survival in larval fishes. The objective of this study was to test this aspect of the TEEL hypothesis by quantifying condition, growth, and trophic level of larval Shortbelly Rockfish Sebastes jordani collected by the California Cooperative Oceanic Fisheries Investigations program and archived at the Ichthyoplankton Collection.

Methods: The trophic level on larval Shortbelly Rockfish was assessed with compound-­specific isotopic analysis of amino acids. Their size at age and survival were estimated with otolith microstructure. Their diet was examined through stomach content analysis.

Result: Observations indicate that larvae consuming prey at a lower trophic level have greater body weight and exhibit faster growth rates. However, feeding at a lower trophic level did not influence body length. The ingested prey responsible for the lower trophic level within larval rockfish could not be determined.

Conclusion: Larval Shortbelly Rockfish consuming prey at a lower trophic level garnered greater body weight and exhibited faster growth rates and provides support for the TEEL hypothesis. However, further research is needed to identify the preferred prey(s) responsible for the more efficient energy transfer.”

Workshop on J. Joel Alicea, “Originalism, the Administrative State, and the Clash of Political Theories”

J. Joel Alicea, Originalism, the Administrative State, and the Clash of Political Theories.

Solum’s Download of the Week for May 3, 2025. Available on SSRN.

This is a synthetic academic workshop generated using enTalkenator (a variation of the Workshop template, using Claude 3.7 Sonnet).

Abstract: “The administrative state is not just premised on a view about what the Constitution means; it is premised on a view about the nature of politics and, in turn, the nature of human beings. The forceful pushback against the administrative state that we have seen unfold over the last three months represents a quite different view about what the Constitution means, and it therefore rests on quite different premises about politics and human nature. In short, originalists’ rejection of the administrative state is not just the result of a disagreement with living constitutionalism’s understanding of our fundamental positive law; it is also a disagreement with the political theory undergirding living constitutionalism and the administrative state.

In laying out that argument, I will proceed in three parts. First, I will show how the main pillars of the administrative state reflect certain progressive politico-theoretical commitments. Second, I will describe an alternative set of politico-theoretical commitments that are often associated with opposition to the administrative state. Finally, I will show how living constitutionalism is the natural extension of the political theory supporting the administrative state and, conversely, how the politico-theoretical commitments opposed to the administrative state find a natural home in originalism as an approach to constitutional theory.

The upshot is that, if President Trump’s transformation of American government succeeds, it will be because the same principles that motivate opposition to the administrative state as a matter of political theory are consistent with the principles that will motivate an originalist Supreme Court to sustain that transformation as a matter of constitutional law. The stakes of this battle over the next few years go beyond a dispute about our positive law; they go to deep questions of political theory.”

Bonus: Two roundtables on a tech interview and … a Buddhist sutta

Here’s something a little different. I gave enTalkenator a text containing (a) Ben Thompson’s excellent interview with Mark Zuckerberg (posted this morning), on, among other things, Meta’s AI strategy and (b) the famous Satipatthana Sutta, the discourse on the four foundations of mindfulness. On reading the interview, I kept thinking of the sutta, and I wanted to see what a Critic Roundtable template would do with these highly contrasting texts.

What you have here are two such conversations. The first was enTalkenated using a version of the Critic Roundtable template that I modified to indicate that the provided text contained more than one article and that the participants should not assume the listeners had read the articles. I used Gemini 2.5 Pro (free version). The output is interesting. It proceeds by focusing first on the interview, then on the sutta, then on both.

The second (beginning immediately after the end of the first at 35:45) uses Claude 3.7 Sonnet. After listening to the first, I made one more modification to the “Setting” in the template — instructing that the participants “explore subtle and sometimes not subtle disagreements among one another” to try to spark more interesting back and forth. I found this conversation much more insightful. The conversation focused on the interview and the sutta together from the beginning. The Gemini version might be more helpful in its initial overview of the texts, especially for someone who has not read them.

Both of these models have produced extremely impressive output for me, in the app and in other contexts. So I don’t present these two roundtables as a head-to-head benchmark. If you enTalkenated this same conversation 100 times, you’d get some crazily interesting conversations and some that are perhaps too straightforward, whichever of these two models you used. But I thought it was interesting and worth sharing. Enjoy!

Interdisciplinary Workshop on “Adversarial testing of global neuronal workspace and integrated information theories of consciousness”

Cogitate Consortium (many authors), Adversarial testing of global neuronal workspace and integrated information theories of consciousness, available at https://www.nature.com/articles/s41586-025-08888-1.

This is a synthetic academic workshop generated using enTalkenator (using an AI-generated interdisciplinary workshop template and authored by Google Gemini 2.5 Pro).

Abstract: “Different theories explain how subjective experience arises from brain activity. These theories have independently accrued evidence, but have not been directly compared. Here we present an open science adversarial collaboration directly juxtaposing integrated information theory (IIT) and global neuronal workspace theory (GNWT) via a theory-neutral consortium. The theory proponents and the consortium developed and preregistered the experimental design, divergent predictions, expected outcomes and interpretation thereof. Human participants (n = 256) viewed suprathreshold stimuli for variable durations while neural activity was measured with functional magnetic resonance imaging, magnetoencephalography and intracranial electroencephalography. We found information about conscious content in visual, ventrotemporal and inferior frontal cortex, with sustained responses in occipital and lateral temporal cortex reflecting stimulus duration, and content-specific synchronization between frontal and early visual areas. These results align with some predictions of IIT and GNWT, while substantially challenging key tenets of both theories. For IIT, a lack of sustained synchronization within the posterior cortex contradicts the claim that network connectivity specifies consciousness. GNWT is challenged by the general lack of ignition at stimulus offset and limited representation of certain conscious dimensions in the prefrontal cortex. These challenges extend to other theories of consciousness that share some of the predictions tested here. Beyond challenging the theories, we present an alternative approach to advance cognitive neuroscience through principled, theory-driven, collaborative research and highlight the need for a quantitative framework for systematic theory testing and building.”

Workshop on Eli Nachmany’s “The Original FTC”

Eli Nachmany, The Original FTC.

Solum’s Download of the Week for April 26, 2025. Available on SSRN.

This is a synthetic academic workshop generated using enTalkenator (a variation of the Workshop template, using Google Gemini 2.5 Pro).

Abstract: “In 1914, Congress established the Federal Trade Commission. At the outset, the agency had a limited set of powers that it exercised in aid of Congress and the courts. Congress provided that the commissioners of the FTC were generally to be free from presidential removal. And in Humphrey’s Executor v. United States, decided in 1935, the Supreme Court upheld these protections as constitutional. The Court in that case rested its holding on the quasi-judicial and quasi-legislative nature of the FTC’s functions. Today, Humphrey’s Executor looms large—its holding suggests that the President cannot remove FTC commissioners at will, even 90 years after the Court decided the case.

That is incorrect. In the years since Humphrey’s Executor, Congress has expanded the FTC’s powers dramatically. It is no longer the quasi-judicial and quasi-legislative agency that the Court in Humphrey’s Executor evaluated. Thus, because the statutory scheme evaluated in Humphrey’s Executor no longer exists as it was when the Court decided the case, the Court’s holding no longer applies to the modern FTC. The difference between the original FTC and the modern FTC compels the following conclusion: The President can remove FTC commissioners at will without contravening Humphrey’s Executor or any other precedent of the Supreme Court.”

A roundtable on three AI articles: Normal tech, biological, or jagged super-intelligence

This episode begins with a brief summary (created by enTalkenator after asking it to generate a summary of multiple, related articles) of three recent articles and then (at 00:14:39) proceeds to a roundtable discussion, generated by modifying the built-in Critics’ Roundtable template. These are the articles discussed:

• Arvind Narayanan & Sayash Kapoor, AI as Normal Technology, https://knightcolumbia.org/content/ai-as-normal-technology.

• Jack Lindsey at al., On the Biology of a Large Language Model, https://transformer-circuits.pub/2025/attribution-graphs/biology.html.

• Ethan Mollick, On Jagged AGI: o3, Gemini 2.5, and Everything After, https://www.oneusefulthing.org/p/on-jagged-agi-o3-gemini-25-and-everything.

Kramer on “Logical Quantification and the Hohfeldian Analysis of Rights”

Matthew H. Kramer, Logical Quantification and the Hohfeldian Analysis of Rights.

Solum’s Download of the Week for April 19, 2025. Available on SSRN.

This is a synthetic academic workshop generated using enTalkenator (a variation of the Workshop template).

Abstract: “Ever since the American legal philosopher Wesley Hohfeld published his analyses of rights in the second decade of the twentieth century, numerous philosophers and legal theorists have grappled with his matrix of legal positions.  Very seldom, however, have the commentators on Hohfeld recognized the importance of logical quantification for any endeavor to elaborate the categories of that matrix fully.  This paper mulls over a few of the manifold ways in which quantification is operative in the legal relationships that Hohfeld delineated.  After furnishing a short conspectus of the Hohfeldian analytical framework which distills those relationships, this paper briefly discusses the two most prominent logical quantifiers and then explores how one’s awareness of the workings of those quantifiers can sharpen one’s grasp of the Hohfeldian analysis.”

Intro to “Functional Connectomics Spanning Multiple Areas of Mouse Visual Cortex”

The MICrONS Consortium, Functional Connectomics Spanning Multiple Areas of Mouse Visual Cortex, available at https://www.nature.com/articles/s41586-025-08790-w.

This is a synthetic introductory class (no expertise assumed!) generated using enTalkenator (selecting Claude 3.7 Sonnet).

Abstract: Understanding the brain requires understanding neurons’ functional responses to the circuit architecture shaping them. Here we introduce the MICrONS functional connectomics dataset with dense calcium imaging of around 75,000 neurons in primary visual cortex (VISp) and higher visual areas (VISrl, VISal and VISlm) in an awake mouse that is viewing natural and synthetic stimuli. These data are co-registered with an electron microscopy reconstruction containing more than 200,000 cells and 0.5 billion synapses. Proofreading of a subset of neurons yielded reconstructions that include complete dendritic trees as well the local and inter-areal axonal projections that map up to thousands of cell-to-cell connections per neuron. Released as an open-access resource, this dataset includes the tools for data retrieval and analysis. Accompanying studies describe its use for comprehensive characterization of cell types, a synaptic level connectivity diagram of a cortical column, and uncovering cell-type-specific inhibitory connectivity that can be linked to gene expression data. Functionally, we identify new computational principles of how information is integrated across visual space, characterize novel types of neuronal invariances and bring structure and function together to uncover a general principle for connectivity between excitatory neurons within and across areas.

Jonathan Choi, “Large Language Models Are Unreliable Judges”

Jonathan H. Choi, Large Language Models Are Unreliable Judges.

Solum’s Download of the Week for April 12, 2025. Available on SSRN.

This is a synthetic academic workshop generated using enTalkenator (a variation of the Workshop template and Claude 3.7 Sonnet).

Abstract: “Can large language models (LLMs) serve as "AI judges" that provide answers to legal questions? I conduct the first series of empirical experiments to systematically test the reliability of LLMs as legal interpreters. I find that LLM judgments are highly sensitive to prompt phrasing, output processing methods, and model training choices, undermining their credibility and creating opportunities for motivated judges to cherry-pick results. I also find that post-training procedures used to create the most popular models can cause LLM assessments to substantially deviate from empirical predictions of language use, casting doubt on claims that LLMs elucidate ordinary meaning.”

Workshop on “Why an overreliance on AI-driven modelling is bad for science” by Arvind Narayanan and Sayash Kapoor

Arvind Narayanan and Sayash Kapoor, Why an overreliance on AI-driven modelling is bad for science, available at https://www.nature.com/articles/d41586-025-01067-2.

This is a synthetic academic workshop generated using enTalkenator (selecting Google Gemini 2.5 Pro Experimental).

The article discussed appeared as a comment in the journal Nature on April 7, 2025, with the subhead: “Without clear protocols to catch errors, artificial intelligence’s growing role in science could do more harm than good.”

Bradley and Siegel, “Revisiting The Essential Functions Thesis: A Structural Account Of The Supreme Court's Roles”

Curtis Bradley and Neil Siegel, Revisiting The Essential Functions Thesis: A Structural Account Of The Supreme Court's Roles.

Solum’s Download of the Week for March 29, 2025. Available on SSRN.

This is a synthetic academic workshop generated using enTalkenator (a variation of the Workshop template and Claude 3.7 Sonnet).

Abstract: “One of the most famous structural arguments in the field of federal courts is the ‘essential functions thesis.’ The thesis, which has been endorsed by a number of prominent scholars and the executive branch, posits that there are implicit, structural limits on Congress's authority to ‘strip’ the Supreme Court of its appellate jurisdiction. Court-stripping, the thesis contends, is not allowed if it would undermine the essential functions of the Court-in particular, maintaining the supremacy and uniformity of federal law. In this Article, we revisit that thesis with three goals in mind. First, we aim to show that the structural arguments underlying the essential functions thesis are relevant to the debate over whether Congress can permissibly ‘pack’ the Court-that is, add Justices to seize ideological control of the institution. Supporters of the essential functions thesis have tended to assume that packing the Court must be constitutionally permissible. In making that assumption, they have mostly resorted to a formalist, textualist frame, a posture that seems puzzling because they do not adopt such a frame in addressing Courtstripping. Second, we contend that the literature has defined the Court's essential functions too narrowly. As we show, the structural rationales that explain why the supremacy and uniformity of federal law are essential functions also suggest other essential functions. Those additional functions, in turn, provide yet another ground for questioning the permissibility of Court-packing. Third, we use the example of the essential functions thesis to invite deeper scholarly consideration of the nature of structural constitutional reasoning and how it should be done, subjects that have received insufficient attention both on the Supreme Court and in the literature. For several reasons, including newfound anxiety over executive branch compliance with adverse judicial decisions, the issues analyzed in this Article are vitally important today.”

Introduction to Littlejohn et al. on “A Streaming Brain-to-Voice Neuroprosthesis to Restore Naturalistic Communication”

Kaylo T. Littlejohn et al. (many authors), A Streaming Brain-to-Voice Neuroprosthesis to Restore Naturalistic Communication, available at https://www.nature.com/articles/s41593-025-01905-6.

This is a synthetic introductory class (no expertise assumed!) generated using enTalkenator (selecting Google Gemini 2.5 Pro Experimental).

Abstract: Natural spoken communication happens instantaneously. Speech delays longer than a few seconds can disrupt the natural flow of conversation. This makes it difficult for individuals with paralysis to participate in meaningful dialogue, potentially leading to feelings of isolation and frustration. Here we used high-density surface recordings of the speech sensorimotor cortex in a clinical trial participant with severe paralysis and anarthria to drive a continuously streaming naturalistic speech synthesizer. We designed and used deep learning recurrent neural network transducer models to achieve online large-vocabulary intelligible fluent speech synthesis personalized to the participant’s preinjury voice with neural decoding in 80-ms increments. Offline, the models demonstrated implicit speech detection capabilities and could continuously decode speech indefinitely, enabling uninterrupted use of the decoder and further increasing speed. Our framework also successfully generalized to other silent-speech interfaces, including single-unit recordings and electromyography. Our findings introduce a speech-neuroprosthetic paradigm to restore naturalistic spoken communication to people with paralysis.

Ramsey on “The Originalist Case Against the Insular Cases”

Michael D. Ramsey, The Originalist Case Against the Insular Cases.

Solum’s Download of the Week for March 29, 2025. Available on SSRN.

This is a synthetic academic workshop generated using enTalkenator.

Abstract: “Concurring in United States v. Vaello Madero, Justice Neil Gorsuch argued that the Insular Cases are contrary to the Constitution’s original meaning and should be overruled. The Supreme Court’s decisions in the Insular Cases, which created a second-class constitutional status for U.S. overseas territories, have also been criticized by leading originalist scholars such as Professors Gary Lawson and Michael Paulsen. However, there is no fully developed scholarly assessment of the Insular Cases from an originalist perspective; their inconsistency with an originalist approach is more assumed than proven. This Article fills that gap. Using the methodology of original public meaning, it considers the constitutional status of U.S. territories from the founding era through the early nineteenth century to the constitutionalization of U.S. citizenship in the Fourteenth Amendment.

Although the matter is somewhat more complicated than Justice Gorsuch’s concurrence may suggest, this Article finds no foundation in traditional originalist sources for the Insular Cases’ differential treatment of overseas territories. To the contrary, it concludes that U.S. territories were widely understood to be broadly encompassed by the Constitution without differentiation until an academic and judicial reassessment at the beginning of the twentieth century, impelled by U.S. acquisition of territories with substantial non-white populations, set the stage for the Court’s newly invented doctrine. This Article thus concludes that Justice Gorsuch’s assessment is correct and should carry weight with the Court’s originalist-oriented majority. Finally, this Article examines from an originalist perspective the implications for territorial government of overruling the Insular Cases, which it concludes would be significant but not substantially destabilizing.”

Academic Workshop on Anthropic Research “Circuit Tracing”

Emmaniel Ameisen et al. (many authors), Circuit Tracing: Revealing Computational Graphs in Language Models, available at https://transformer-circuits.pub/2025/attribution-graphs/methods.html.

This is a synthetic academic workshop generated using enTalkenator (selecting Claude 3.7 Sonnet).

Abstract: We introduce a method to uncover mechanisms underlying behaviors of language models. We produce graph descriptions of the model’s computation on prompts of interest by tracing individual computational steps in a “replacement model”. This replacement model substitutes a more interpretable component (here, a “cross-layer transcoder”) for parts of the underlying model (here, the multi-layer perceptrons) that it is trained to approximate. We develop a suite of visualization and validation tools we use to investigate these “attribution graphs” supporting simple behaviors of an 18-layer language model, and lay the groundwork for a companion paper applying these methods to a frontier model, Claude 3.5 Haiku.

Intro to Anthropic Research “On the Biology of a Large Language Model”

Jack Lindsey et al. (many authors), On the Biology of a Large Language Model, available at https://transformer-circuits.pub/2025/attribution-graphs/biology.html.

This is a synthetic introductory class (no expertise assumed!) generated using enTalkenator (selecting Claude 3.7 Sonnet).

Abstract: We investigate the internal mechanisms used by Claude 3.5 Haiku — Anthropic's lightweight production model — in a variety of contexts, using our circuit tracing methodology.

Intro to Anthropic Research on “Circuit Tracing: Revealing Computational Graphs in Language Models”

Emmaniel Ameisen et al. (many authors), Circuit Tracing: Revealing Computational Graphs in Language Models, available at https://transformer-circuits.pub/2025/attribution-graphs/methods.html.

This is a synthetic introductory class (no expertise assumed!) generated using enTalkenator (selecting Claude 3.7 Sonnet).

Abstract: We introduce a method to uncover mechanisms underlying behaviors of language models. We produce graph descriptions of the model’s computation on prompts of interest by tracing individual computational steps in a “replacement model”. This replacement model substitutes a more interpretable component (here, a “cross-layer transcoder”) for parts of the underlying model (here, the multi-layer perceptrons) that it is trained to approximate. We develop a suite of visualization and validation tools we use to investigate these “attribution graphs” supporting simple behaviors of an 18-layer language model, and lay the groundwork for a companion paper applying these methods to a frontier model, Claude 3.5 Haiku.

Kerr on “Data Scanning and the Fourth Amendment”

Orin S. Kerr, Data Scanning and the Fourth Amendment.

Solum’s Download of the Week for March 23, 2025. Available on SSRN.

This is a synthetic academic workshop generated using enTalkenator.

Abstract: “A crucial question of Fourth Amendment law has recently divided courts: When government agents conduct a digital scan through a massive database, how much of a "search" occurs? The issue pops up in contexts ranging from geofence warrants and reverse keyword searches to the installation of Internet pen registers. When a government agent runs a filter through a massive database, resulting in a list of hits, is the scale of the search determined by the size of the database, the filter setting, or the filter output? Fourth Amendment law is closely attuned to the scale of a search. No search means no Fourth Amendment oversight, small searches ordinarily require warrants, and limitless searches are categorically unconstitutional. But how broad is a data scan? 

This essay argues that that Fourth Amendment implications of data scans should be measured primarily by filter settings. Whether a search occurs, and how far it extends, should be based on what information is exposed to human observation. This standard demands a contextual analysis of what the output reveals about the dataset based on the filter setting. Data that passes through a filter is searched or not searched depending on whether the filter is set to expose that specific information. The proper question is what information is expressly or implicitly exposed, not what raw data passes through the filter or the raw data output. The implications of this approach are then evaluated for a range of important applications, among them geofence warrants, reverse keyword searches, and Internet pen registers.”