New APPS

Art, Politics, Philosophy, Science. Mostly Philosophy.

recent posts

about

    • Phenomenological seams (part 2): Borgmann, Heidegger and Derrida

      Last time, I set up a topic by reading Brett Frischmann’s and Paul Ohm’s “Governance Seams.”  Governance seams are frictions and inefficiencies that can be designed into technological systems for policy ends.  In this regard, “Governance seams maintain separation and mediate interactions among components of sociotechnical systems and between different parties and contexts” (1117).  Here I want to suggest that governance seams have a very close relation to phenomenological ones.  To get there, let me take a detour into an older philosophy of technology paper, Albert Borgmann’s “Moral Significance of the Material Culture.”  Borgmann is concerned with what he takes to be the way that moral and ethical theory ignore material culture, whether they emphasize theory or practice.   Via a paper by Csikszentmihalyi and Rochberg-Halton, he arrives at a distinction between things that he calls “commanding” and “disposable.”  The moral complaint is about the “decline of commanding and the prominence of disposable reality” (294).  Following them, Borgmann distinguishes between a musical instrument and a stereo.

      “A traditional musical instrument is surely a commanding thing,” he writes:

      “It is such simply as a physical entity, finely crafted of wood or metal, embodying centuries of development and refinement, sometime showing the very traces of its service to many generations. An instrument particularly commands the attention of the student who, unless she is a prodigy, must through endless and painstaking practice adjust her body to the exacting requirements of this eminently sensitive thing.” (294)

      After some more similar description, emphasizing the multisensory experience of witnessing someone play an instrument, he turns to the stereo.  Certainly a “stereo produces music as well or, in fact, much better” and some stereos are big.  Nonetheless, “as a thing to be operated, a stereo is certainly not demanding. Nor do we feel indebted to its presence the way we do when we listen to a musician.  We respect a musician, we own a stereo” (295).  The stereo is on the rise, perhaps because “the history of the technology of recorded music is the history of obliging ever more fully the complaint about the burden and confinement of live music,” or “more positively” it is a “promise to provide music freely and abundantly” which is tied to “the promise of general liberty and prosperity – the promise that inaugurated the modern era” (295).

      (more…)

    • Phenomenological Seams (part 1)

      In a recent paper, Brett Frischmann and Paul Ohm introduce the idea of “governance seams,” which are frictions and inefficiencies that can be designed into technological systems for policy ends.  In this regard, “Governance seams maintain separation and mediate interactions among components of sociotechnical systems and between different parties and contexts” (1117).  Their first example is a university’s procedure for anonymous exams.  There, a number of different friction points are added to make sure that professors do not know whose exams they are grading: students receive unique identifiers from the registrar and use only these on the exams; they type but do not write answers; once the exams are scored, the registrar matches the numbers back with individual student names; and so forth.  The professor will likely not be in the room during the exam, so the university will have to provide a neutral proctor.  The exam might also take place in a specified location.  There will also be rules and penalties designed to ensure that none of the seams are crossed without permission.  Together, these governance seams design the system for fairness, or at least to eliminate one potential source of bias in grading.  They’re also pretty inefficient in that they require a bunch of resources be allocated to them, but places with that sort of anonymous grading figure it’s worth it for the fairness bump.

      As the paper goes on to argue, such governance seams are ubiquitous and important, because they enable us to design sociotechnical systems to facilitate certain outcomes that might otherwise not happen.  That is, seams open a space for governance:

      (more…)

    • Privacy and “Injury in Fact,” Part 6: Defamation

      As a final installment of reviewing some older “injury in fact” cases, I’d like to look at a few older state libel cases, because the distinction emerges especially clearly in them.  A North Carolina case, for example, noted that “he who publishes slanderous words even as those of a third person with the intent, (to be collected from the mode, extent and circumstances of the publication,) that the charges should be believed, does an injury in fact to the person slandered and ought to answer for it” (Hampton v. Wilson, 15 N.C. 468, 470 (1834).  Here’s a few cases in more detail.  The 19c gender politics is really helpful in seeing how their minds worked on defamation per se.

       

      (a) Chastity in Iowa

      A pair of Iowa cases are particularly clear.  In Abrams v. Foshee, the Court was asked to rule whether accusing a woman of having an abortion was actionable as slander.  The Court lays out its reasoning particularly clearly:

      “To maintain an action of slander, the consequence of the words spoken, must be to occasion some injury or loss to the plaintiff, either in law or fact. As the declaration in this case, claims no special damages, or a loss or injury, in fact, we are left to inquire whether the charges referred to in the instructions refused, was of such a character as to amount to an injury in law. To determine this, it becomes material to ascertain in what cases this action may be maintained, without proof of special damages. Starkie, in his work on Slander, page 9, lays down the rule, that such action may be maintained "when a person is charged with the commission of a crime; when an infectious disorder is imputed; and when the imputation affects the plaintiff in his office, profession, or business." In this case, we only need examine the rule so far as it relates to the charge of a crime. And what is that rule? In Cox and wife v. Bunker and wife, Morris, 269, the Supreme Court of this territory, recognized the rule laid down in Miller v. Parish, 25 Mass. 384, 8 Pick. 384, as the proper one. And in that case it is said, that " whenever an offense is charged, which if proved, may subject the party to a punishment, though not ignominious, but which brings disgrace upon the party falsely accused, such an accusation is actionable. And this is, perhaps, as correct, and at the same time as brief a statement of the general rule, as has been given. For while the rule is variously stated, by different authors and judges, yet in all of them, it is laid down as necessary that the charge shall impute a punishable offense.” (Abrams v. Foshee, 3 Iowa 274, 277-8 (1856)).

      That is, if the false statement would have subjected the victim to legal punishment if true, it was considered libel per se – actionable as an act, independent of any damages sustained.  In 1843, “willful killing of an unborn quick child, by an injury, etc., was made manslaughter” (278).  This statute was repealed in 1851. So abortion was not a crime. Plaintiffs urged that the fetus was a “human being” and thus subject to murder.  The Court, at length, disagreed, citing both statute and common law precedents (including Coke and Blackstone) to the effect that abortion was not “murder” even if it were a misdemeanor or otherwise bad.

      (more…)

    • Tariff policy by chatbot?

      I desperately and truly wish that I'd made this up.  Alas, the Verge reports:

      "Economist James Surowiecki quickly reverse-engineered a possible explanation for the tariff pricing. He found you could recreate each of the White House’s numbers by simply taking a given country’s trade deficit with the US and dividing it by their total exports to the US. Halve that number, and you get a ready-to-use “discounted reciprocal tariff.” The White House objected to this claim and published the formula it says that it used, but as Politico points out, the formula looks like a dressed-up version of Surowiecki’s method. In case you weren’t sure, Surowiecki calls this approach “extraordinary nonsense.” So why did Trump’s team use it? Well, like plenty of people who’ve realized their homework is due in three hours’ time, it seems like they may have been tempted by AI."

      Wait, what?

      "A number of X users have realized that if you ask ChatGPT, Gemini, Claude, or Grok for an “easy” way to solve trade deficits and put the US on “an even playing field”, they’ll give you a version of this “deficit divided by exports” formula with remarkable consistency. The Verge tested this with the phrasing used in those posts, as well as a question based more closely on the government’s language, asking chatbots for “an easy way for the US to calculate tariffs that should be imposed on other countries to balance bilateral trade deficits between the US and each of its trading partners, with the goal of driving bilateral trade deficits to zero.” All four platforms gave us the same fundamental suggestion.

      There is some variation. Grok and Claude specifically suggested halving the tariff figure to generate what Grok calls a “reasonable” result, much like Trump’s “discount” idea. Ask for a 10 percent baseline tariff and the systems also disagree on whether that should be added to the total tariff rate or not. But answers from across the four chatbots have more similarities than differences.

      As I write this, the Dow Jones is down 3.98%. 

    • Privacy and “Injury in Fact,” Part 5: Early State cases

      I’ve been indirectly pursuing the question of the problems faced by privacy plaintiffs in data cases by looking at the origins of the Supreme Court’s standing doctrine.  Basically, plaintiffs have to show an “injury in fact,” and courts often find privacy harms not to meet this standard.  Although presented as dating from time immemorial, the injury in fact requirement was actually announced rather abruptly in 1970 (all of this is part 1).  I’ve been exploring the historical antecedents that will help understand what that language implies – in a very early Supreme Court case (part 2), in other federal case law (part 3), and in federal cases about the Administrative Procedure Act (part 4).  Here I want to extend the genealogy into some early state cases; I’ll draw a somewhat arbitrary cutoff at 1930.  This time I’ll look at a general potpourri of cases. Next time I want to specifically look at a few libel cases because the language is especially clear in them.  I don’t claim this to be exhaustive (and I’m ignoring some of the cases around trusts and deeds because the facts in them are often very confusing, but I think it collectively paints a pretty good picture of what “injury in fact” connoted in Data Processing.

      On the whole, the cases point to the legal vs non-legal harm distinction I’ve been developing, As the New Jersey Supreme Court used the concept in an estate case, “there was no injury, in fact or in contemplation of law, to prevent in this case the merger” of the estates (Den ex dem. Wills v. Cooper, 25 N.J.L. 137, 159 (1855).

      (more…)

    • AI is (still) not a (copyright) Author: The Thaler appellate decision

      AI Copyright sad chatbotThe Federal Circuit has affirmed the denial of copyright protection to an AI-generated image on the grounds that copyright requires a human author.  As far as I know this was the expected outcome; I certainly think it’s correct.  I talked about the case a bit and made a couple of policy arguments against AI copyright here, when the lower-court ruling came out.

      The appellate decision lists several reasons AI cannot be an author: (1) copyright authorship is premised on the capacity to hold property, which AI cannot; (2) copyright duration is tied to the author’s lifespan; (3) copyright includes inheritance conditions, and machines don’t have heirs; (4) Copyright transfer requires a signature, but “machines lack signatures, as well as the legal capacity to provide an authenticating signature;” (5) authors are protected regardless of their “nationality or domicile,” but machines have neither; (6) authors have intentions whereas “Machines lack minds and do not intend anything;” (7) when the copyright act does talk about machines, it always talks about them as tools.

      As the court summarizes:

      “All of these statutory provisions collectively identify an “author” as a human being. Machines do not have property, traditional human lifespans, family members, domiciles, nationalities, mentes reae, or signatures. By contrast, reading the Copyright Act to require human authorship comports with the statute’s text, structure, and design because humans have all the attributes the Copyright Act treats authors as possessing. The human-authorship requirement, in short, eliminates the need to pound a square peg into a textual round hole by attributing unprecedented and mismatched meanings to common words in the Copyright Act.” (12)

      (more…)

    • The Prisoner’s Dilemma of AI Cooperation

      In a recent piece on Lawfare, Simon Goldstein and Peter N. Salib make the case that AI with cooperation is better than attempting some sort of AI race, unlike what virtually all of the relevant policymakers in the US advocate.  Thus, in response to the Chinese DeepSeek model, US policymakers are doubling down on the idea that the US must “dominate” AI and win against its geopolitical rival.  Goldstein and Salib write:

      “In any high-stakes competition to obtain powerful military technology, the closer the game, the more sense it makes to declare a truce and cooperate. Cooperation can help to ensure that both superpowers obtain transformative AI around the same time. This preserves the current balance of power, rather than unsettling it and inviting extreme downside risks for both nations.”

      (more…)

    • Privacy and “Injury in Fact,” Part 4: Administrative Procedure Act Cases

      Here I want to complete my review of federal legal precedents for the Supreme Court’s sudden invocation of “injury in fact” language to understand judicial standing in its 1970 Data Processing decision (recall the earlier installments: first, second, third. The first one explains the issue; if you want to escape my rummaging through the archive, you can skip to this one).  Congress passed the Administrative Procedure Act in 1946 to, well, regulate administrative procedures and provide checks against their being arbitrary (this is one of the Acts that virtually of Trumps recent executive orders violates).  Litigation about agency actions after the APA thus had to route through the APA, which imposed its own standards for judicial review of agency actions.

      Here, the language of the lower courts gets very close to the issues in Data Processing.  Consider first Curran v. Laird, in which a maritime union sought enforcement of the Cargo Preference Act, which required that American ships be used for military cargo.  After awarding standing, the DC Cicruit concluded that the decision was a matter for agency discretion under the APA, ruling in favor of the government on the merits.  The Court opens its standing discussion by noting that “plainly [plaintiffs are] aggrieved in fact by the allegedly unlawful action of the Secretary of Defense.” The Court then writes that:

      (more…)

    • Injury in fact Part 3: Caselaw after Auld

      By Gordon Hull

      I’ve been pursuing (first, second) what it means for standing law – basically, the determination that someone has a case that can be addressed by an Article III court – to require that plaintiffs show an “injury in fact,” a requirement that emerged suddenly in the Supreme Court’s Data Processing decision in 1970.  The requirement is at considerable odds with the caselaw before it, and it has puzzled commentators.  Last time, I looked at the Supreme Court’s first use of the term, an early 19th Century case called Hepburn and Dundas v. Auld.  There I suggested that the court’s application of the term to a contract – looking for whether someone suffered harm, outside of the bare violation of a contract terms – wasn’t unlike what the Data Processing court seemed to be doing, in arguing that plaintiffs could demonstrate either legal injury (because of statutory violation) or harm of some other sort (injury in fact).  Here I want to dig into some of the earlier caselaw; what I think that caselaw establishes is that the concept of “injury in fact” generally works to emphasize material harm, as opposed to some sort of statutory or “merely” legal injury.  That sort of contrastive usage isn’t enormously common, but I do think it’s pretty consistent.

      There’s three areas where you could talk about this – older cases, regulatory cases, and those relating to the Administrative Procedure Act (APA).  I’ll look at the first two this time, and divide them into Supreme Court and lower court decisions.

      (more…)

    • Explainable AI as Disciplinary Power

      I want to take a break from judicial standing doctrine to note a recent and helpful paper by Emily Sullivan and Atoosa Kasirzadeh about explainable AI.  Explainable AI is a research agenda – there’s a lot of papers and techniques (for a current lit review, see here) – that is designed to get at a central problem in using AI: we often have no idea why machine learning systems produce the outputs they do.  In a variety of contexts, ranging from safety critical systems to democratic governance, being able to understand why the algorithm made the prediction is did is important. Hence the research agenda. 

      First, a little detour. Algorithmic governance can be disciplinary in that it can nudge people inexorably toward conforming with norms, whether social or statistical.  Insurance has been well-studied for its normalizing techniques.  In an early paper on privacy unraveling, Scott Peppett showed how the addition of smart-driving surveillance (where insurers give a discount to people who install these devices that record their speed, when they drive, etc.) generate a downward ratcheting on privacy: users who are good drivers have the incentive to adopt the devices, since they get lower insurance rates.  Those who are in the next tier down (above-average drivers) have an incentive to get the devices because that associates them with the good drivers.  And so it goes, until only the worst drivers are declining the surveillance.  And at some point, not having the surveillance device becomes a stigma that raises your rates.  So pretty soon, surveillance devices can become normal.  In the meantime, once drivers have the surveillance installed, surveillance-enabled insurance can nudge them to drive less at night and to otherwise comply with whatever the insurance company says makes you a good risk.  All of that can be automated – the insurance app can tell you, real time, how your driving is impacting your premium.

      (more…)