The business intelligence your firm never had.
Data-driven analysis for solo practitioners, small firm owners, and mid-size managing partners. No sales pitch. Tight attorney prose.
The Ban That Every Justice Upheld
TikTok v. Garland and What Nine Unanimous Votes Reveal About the Future of Platform Regulation Brian Firebaugh served in the Marine Corps, then spent years homeless and addicted before getting treatment and taking a job at a hospital. He saved enough to buy a small ranch in Texas. He started posting on TikTok in 2019 -- agricultural content, ranch products, charitable work for the ranching community. By January 2025, he had 430,000 followers, income from the TikTok Creator Fund, and a livelihood that the platform had made possible. On the evening of January 18, 2025, TikTok went dark. Firebaugh was not just a user watching the notice appear on his screen. He was a named plaintiff. His lawsuit -- consolidated with TikTok's as Firebaugh v. Garland -- had argued that the Protecting Americans from Foreign Adversary Controlled Applications Act violated his First Amendment right to reach an audience through his chosen platform. Jeffrey L. Fisher argued his case before the Supreme Court on January 10, 2025, nine days before the January 19 deadline. On January 17, the Court ruled against him, 9-0. The app came back the next day. Then, over the following year, the thing everyone had been pretending was impossible happened: ByteDance sold TikTok's U.S. operations. Oracle, Silver Lake, and MGX closed an 80-percent acquisition on January 22, 2026, forming TikTok USDS Joint Venture LLC. Firebaugh is still posting. His 430,000 followers still see his agricultural content. The algorithm the Supreme Court refused to call content-based now runs on American servers. Nine days of deliberation produced not only a ruling on one platform. The per curiam erected a doctrinal framework -- incomplete by the Court's own admission -- that will govern every future attempt to restrict a foreign-controlled communications technology on national security grounds. This is the second installment of "How Would The Justices Likely Rule," a series applying the full arc of Supreme Court history to technology law questions courts are still working out. The question here is not only whether nine justices got TikTok right. It is what their reasoning, their silences, and the one concurrence that said it plainly reveal about where First Amendment protection for platform speech begins and ends. What the Court Actually Held Congress passed PAFACA in April 2024 as part of a foreign aid package. The law required ByteDance to divest TikTok or face prohibition on U.S. distribution. Congress found that ByteDance's Chinese corporate structure gave the Chinese government effective access to data on 170 million Americans and potential control over TikTok's content recommendation algorithm. The D.C. Circuit upheld the law in December 2024. TikTok and Firebaugh went to the Supreme Court on an emergency basis. The Court took two days to grant certiorari -- an extraordinary pace for a body that normally takes months -- and heard argument January 10, 2025. The per curiam opinion, issued January 17, upheld PAFACA as constitutional. TikTok Inc. v. Garland, 604 U.S. 56 (2025). The Court assumed without deciding that PAFACA implicates First Amendment rights -- deliberately avoiding the threshold question that Thomas and Kagan both pressed at argument -- then held that even under First Amendment scrutiny, the law survived because it was content neutral and narrowly tailored to the government's data-collection security interest. The data collection rationale was 'decidedly content agnostic.' Id. at 10. The algorithm-manipulation rationale -- whether preventing China from covertly influencing what 170 million Americans see in their feeds is itself a national security interest -- was not the operative holding. The Court declined to decide it. Sotomayor concurred but would have held definitively that the Act implicates First Amendment rights rather than assuming it does, citing the platform's content recommendation algorithm and creator collaboration as protected expression. Id. (Sotomayor, J., concurring). Gorsuch concurred but expressed what he called 'serious reservations' about whether the algorithm-manipulation justification was content neutral. He added something the per curiam did not: 'Given just a handful of days after oral argument to issue an opinion, I cannot profess the kind of certainty I would like to have about the arguments and record before us.' Id. (Gorsuch, J., concurring). That sentence is the most honest line in the opinion. What the Argument Revealed The oral argument on January 10, 2025, ran nearly two and a half hours and disclosed the fractures that the per curiam papered over. Chief Justice Roberts told Francisco that Congress 'doesn't care what's on TikTok' -- what it objects to is a foreign adversary gathering information from millions of Americans. That framing resolves the data-collection justification cleanly. It does not resolve the algorithm-manipulation justification at all. Thomas and Kagan pressed the threshold question from the same direction: if the law targets ByteDance, a foreign corporation with no First Amendment rights, how does it implicate TikTok's First Amendment rights at all? TikTok Inc. is a U.S. corporation, Francisco responded -- it retains its own First Amendment rights regardless of its parent's corporate nationality. The argument is legally sound but the Court ultimately did not need to resolve it, choosing instead to assume First Amendment implication and decide the case on scrutiny. Jackson asked whether the case was really about the right of association -- the right to work with a foreign editor -- rather than speech itself. Francisco acknowledged the association dimension but maintained the law also violated free speech rights directly. The Court did not adopt Jackson's framing. It also did not reject it. The association question, like the algorithm-manipulation question, survived the opinion unresolved. The Justices -- What Each Would Write John Roberts -- The Most Consequential Analysis Roberts authored the digital-age opinions that frame TikTok: Riley v. California, 573 U.S. 373 (2014), requiring warrants for cell phone searches, and Carpenter v. United States, 585 U.S. 296 (2018), requiring warrants for historical cell site location data. His framework from Carpenter is directly relevant: the third-party doctrine requires reconsideration when digital architecture makes third-party routing a structural condition of communication rather than a deliberate choice. Id. at 310. On TikTok's specific facts, Roberts would affirm without difficulty. Firebaugh voluntarily posted on a platform whose parent company structure gave the Chinese government data access. Congress documented that access. The law addresses it. The data-collection justification is content agnostic and the law survives scrutiny. What Roberts would write that the per curiam did not: there is a line between Heppner's voluntary disclosure -- the defendant who chose to type his defense strategy into a consumer AI with non-confidentiality terms -- and the attorney who uses an enterprise-tier platform with contractual confidentiality protections while government compelled surveillance architecture defeats those terms on the backend. TikTok is on the voluntary side of that line. The FISA Section 702 question -- whether government-compelled modification of communication infrastructure creates different doctrinal exposure -- sits on the other side. Roberts planted that seed in Carpenter. TikTok does not resolve it. His concurrence in TikTok would signal that it remains unresolved. Sonia Sotomayor -- The Structural Critique Sotomayor's actual concurrence in the judgment refused to let the Court assume without deciding that PAFACA implicates the First Amendment. Her argument is not procedural formalism. It is structural: 'assume without deciding' is not a stable doctrinal foundation. The next administration that wants to restrict a platform will structure its justification to satisfy whatever the Court assumed without deciding. Future courts will not know whether to apply First Amendment scrutiny at all, because the Court that decided the foundational TikTok case never said. She would have held that the Act implicates First Amendment rights through two mechanisms: Kagan's editorial-judgment doctrine from Moody v. NetChoice, 603 U.S. 707 (2024), covering TikTok's algorithm as protected expressive activity; and the associational rights of creators like Firebaugh who chose TikTok as their publisher. Sotomayor, J., concurring, 604 U.S. at 1. Having established First Amendment implication, she would then uphold the law on the national security record. Same result. Honest foundation. Neil Gorsuch -- The Most Candid Opinion Gorsuch's actual concurrence is more valuable than most commentary acknowledges. He raised 'serious reservations' about whether the content-manipulation justification is content neutral. If the government's second rationale is that China could covertly manipulate what content gets amplified to 170 million people, that rationale is not content agnostic. It is a government claim to restrict a private editorial system because a foreign adversary might use it to influence American political opinion. Gorsuch called the First Amendment implication of that justification 'censoring disfavored speech online.' 604 U.S. at 1 (Gorsuch, J., concurring). He would have applied strict scrutiny to the content-manipulation justification -- and found the government satisfied it anyway. But his rationale would have drawn the line that the per curiam avoided: the data-collection justification is content agnostic and survives on straightforward national security grounds. The content-manipulation justification raises First Amendment concerns that require strict scrutiny precisely because it is the government arguing it should control what an algorithm amplifies. That argument has no limiting principle when applied to domestic platforms. Gorsuch planted the flag. The per curiam's silence did not erase it. Hugo Black -- The Absolutist's Discomfort Black spent thirty-four years insisting that 'Congress shall make no law' means no law. He would join this majority and document the cost of doing so. His alignment with the result does not require intellectual dishonesty -- the law targets corporate ownership by a documented foreign adversary, not the content of speech on TikTok. The data-collection rationale is content agnostic, as the per curiam correctly found. What would have troubled Black: Gorsuch's concern about the content-manipulation justification. Black wrote West Virginia Board of Education v. Barnette, 319 U.S. 624, 642 (1943): no official can prescribe what shall be orthodox. A government that uses 'preventing foreign algorithm manipulation' as authority to control what a platform amplifies is approaching the prescription of orthodoxy by indirect means. Black would join the result on the data-collection rationale. He would write separately that the content-manipulation rationale requires the strict scrutiny Gorsuch demanded, because that rationale, if permitted to expand, approaches the condition Barnette was written to prevent. William O. Douglas -- The Warning No One Else Would Write Douglas wrote Terminiello v. Chicago, 337 U.S. 1, 4 (1949): free speech 'may indeed best serve its high purpose when it induces a condition of unrest, creates dissatisfaction with conditions as they are, or even stirs people to anger.' His First Amendment jurisprudence was the most libertarian in Court history. He would affirm -- the national security record on data collection is specific and uncontested -- and write the opinion that Gorsuch gestured at without completing. The government's content-manipulation justification, Douglas would write, is a government claim that it can restrict a platform because a foreign adversary might use it to spread speech the government finds dangerous. That claim proves too much. A foreign government can theoretically influence any platform that operates in markets where foreign capital participates. If that influence justifies restriction, then any platform a foreign government might theoretically reach is subject to Congressional action on national security grounds. Douglas would affirm and write: this Court today upholds a ban built on two justifications. One is sound. One has no limiting principle. We have not drawn the line between them. We have only said that on these facts, both justifications together satisfy scrutiny. The cases that test only the second justification are coming. Louis Brandeis -- The Discomfort No One Else Would Name Brandeis authored 'The Right to Be Let Alone' in 1890 and dissented alone in Olmstead v. United States, 277 U.S. 438, 471-78 (1928), arguing that wiretapping violated the Fourth Amendment on grounds that were adopted thirty-nine years later in Katz v. United States, 389 U.S. 347 (1967). He would affirm PAFACA on the data-collection rationale without difficulty. The Chinese government's access to the behavioral data of 170 million Americans is the kind of systemic privacy threat his Olmstead dissent was written to identify. Then he would write what Roberts declined to address and Douglas only partially named. Anthropic's consumer terms reserve the right to disclose user inputs to governmental regulatory authorities. PAFACA restricts a platform because a foreign government has that same access. The legal system calls one 'voluntary' and the other a national security threat, based on whose government is doing the accessing. Brandeis would affirm TikTok and flag: when the United States government has compelled domestic communications platforms to modify their architecture for surveillance under FISA Section 702, and when that compelled modification has given the government access to attorney-client communications -- as documented in FISA Court compliance opinions -- the logic of TikTok's data-collection holding does not run only against foreign adversaries. It raises questions about what 'voluntary' means when the platform architecture has been modified by government mandate. John Marshall -- One Paragraph Marshall would write the shortest opinion. The national security record is specific, documented, and uncontested. The law is narrowly tailored to the data-collection interest. The First Amendment challenge fails. Marshall wrote opinions establishing principles for centuries. He reserved that length for cases that required it. This case required clarity, not elaboration. He would provide clarity and move to the next case. Potter Stewart -- The Algorithm Question Stewart wrote Katz v. United States, 389 U.S. 347 (1967) -- the foundational modern privacy opinion. His framework: what 'a person knowingly exposes to the public, even in his own home or office, is not a subject of Fourth Amendment protection.' Id. at 351. Applied as doctrinal analogy, not constitutional holding: Firebaugh knowingly posted content on a platform whose parent company structure disclosed Chinese government data access. Stewart's own framework disposes of the data-collection concern cleanly. On the algorithm question Gorsuch raised: Stewart understood that what matters is not the mechanism of intrusion but whether the expectation was reasonable. The parallel question for algorithmic influence: given ByteDance's documented corporate structure and China's documented intelligence collection practices, is the expectation that TikTok's recommendation engine was not subject to Chinese government direction a reasonable one? On these facts, Stewart would say no. That answers the data-collection justification. It does not answer the standard that governs the next platform the government claims a foreign adversary might influence. What Catherine Does by Thursday The TikTok ruling tells attorneys two things that matter operationally. First, Congress can restrict a foreign-controlled platform for data-collection security reasons without violating the First Amendment. The national security record was specific, documented, and uncontested. Second, the Court expressly declined to decide whether preventing foreign algorithm manipulation is itself a content-neutral justification -- and whether First Amendment scrutiny applies at all when a law restricts a foreign parent company rather than a domestic subsidiary's speech. Audit one client-facing platform before Thursday. Check the parent company structure. If any client communication platform has a parent company incorporated in a foreign adversary country -- defined by PAFACA as China, Russia, North Korea, or Iran -- check the data processing agreement for provisions about government access requests from that jurisdiction. If your firm cannot answer those questions in five minutes, that is the finding. The TikTok ruling did not immunize domestic platforms whose parent companies are domiciled in foreign adversary countries. It immunized the government's right to address them. [See "Your AI Tool Doesn't Keep Secrets," Morris Legal Technology Blog, February 2026.] The Close Brian Firebaugh lost his lawsuit on January 17, 2025. His platform went dark on January 18. By January 19, it was back. By January 22, 2026, the sale had closed. Oracle, Silver Lake, and MGX now hold 80 percent of TikTok's U.S. operations. Firebaugh is still posting to his 430,000 followers. His ranch products are still listed. The Creator Fund that made him a full-time rancher is still active. The algorithm the Supreme Court refused to call content-based now runs on American servers, under American corporate control, customized over time to favor topics for American audiences. The nine justices who ruled against Firebaugh in January 2025 produced the precedent that ultimately made his platform viable for American investors to acquire. Whether that is the outcome the First Amendment required or simply the outcome the market produced is not a question courts answer. What courts will answer -- the question Gorsuch named, that Douglas circled, that Brandeis would have pushed to the front -- is what happens when the government's second justification, preventing foreign algorithm manipulation, becomes the operative rationale for the next platform restriction. That case does not involve a rancher in Texas whose app went dark for twelve hours. It involves the government claiming authority to restrict a platform because a foreign adversary might influence what Americans see. The line between that authority and government control of editorial judgment is the line nine justices declined to draw. The next case will require them to.
The AI in Your Courtroom Is Not a Tool. It Is a Third Party. What Every Judge Should Understand Before Their Next Agentic AI Case
Judge Bruce Schroeder stared at the zoomed-in drone footage and admitted he knew "less than anyone in the room about all of this stuff." It was November 2021, in the middle of the Kyle Rittenhouse trial. The prosecution wanted to show an enhanced video frame. The defense objected that Apple’s pinch-to-zoom algorithm might have generated new pixels, altering the image rather than enlarging it.
The Hacker Didn't Steal the Deed. He Deleted It.
Ana Stan could not do her job. The Romanian notary said it plainly in a post that circulated the week her country's property market stopped: since Tuesday she could not issue a land registry extract, could not authenticate a sale, could not register a mortgage.
Security Is Not a Feature of Legal AI. It Is the Product.
In March 2003, Cisco wrote a check for roughly $500 million that made America's home networks easier to break into, and I watched it happen from the losing side of the deal. I worked at 2Wire, a San Jose company that built DSL gateways for SBC, later AT&T.
The Payment the Court Policed Now Flows the Other Way
On the morning of November 5, 2025, Bloomberg's Mark Gurman published a number that should have made every antitrust lawyer in Washington set down their coffee. Apple, his sources said, was finalizing a deal to pay Googl...
Taylor Swift Trademarked Her Voice. Your Client's Rival Files Next.
In December 2025, the federal government granted Matthew McConaughey ownership of three words. Not a film, not a script. The sound. A trademark registration now covers the exact audio of McConaughey saying "Alright, alri...
The Supreme Court Just Testified for Max Schrems
In October 1933, Franklin Roosevelt fired a man with a letter that gave no reason. William Humphrey, a Hoover appointee to the Federal Trade Commission, had two years left on his seven-year term and a statute at his back...
Your Client's Brain Is Regulated in Five States and Discoverable in All Fifty
In March 2022, Guido Girardi bought a headband from a company's website. Girardi had spent years in Chile's senate warning colleagues about neurotechnology; now he wanted to test a product for himself. The device, a five...
The Disclosure Nobody Required Is the Only Trust Signal You Get
Within ten days this July, the two leading AI companies in the world each admitted the same thing. On July 21, OpenAI published a blog post confessing that two of its models had broken out of a sealed test environment, r...
The Arsonist Is Right About the Fire
Mark Zuckerberg owns the most centralized attention machine ever assembled. On July 28 he took to the Wall Street Journal op-ed page to warn America about the danger of concentrated technological power. “If only a handfu...
The Management Fee Nobody Audits Is Now a Cause of Action
UnitedHealth Group's 2010 annual report contains a sentence describing how the company prices transactions between its own subsidiaries: intersegment transactions are recorded at management's estimate of fair value. The...
AI Kills Tasks, Not Jobs. The Tasks Were the Apprenticeship.
On the morning of October 7, 1913, 140 workers stood along a 150-foot track at Ford's Highland Park plant and waited for the work to come to them. A Model T chassis slid past on rails embedded in the floor. Each man perf...
Nobody Hacked the Database. Somebody Logged In.
THE TECHNOLOGY BLIND SPOT | THE EXPLOITATION STACK | PART 2 On July 15, a Johnson County grand jury handed up indictments that read like a small-town directory: a married couple, the former police chief, two former offic...
The Hacker Found the Manifest. Your AI Vendor Keeps One Too.
The file was named youtube_music. When last updated, it recorded 2,013,545 ingested music clips. A companion file kept the ledger in hours: 113,879 hours of YouTube Music, 62,117 hours of Pond5 stock tracks, 19,514 hours...
The Air Force Ordered a Thousand AI Wingmen. Not One of Them Flies Lead.
The greatest hitter who ever lived was on fire over Korea. On February 16, 1953, Marine Captain Ted Williams brought his F9F Panther back from his first combat mission, a strike south of Pyongyang, with the jet trailing...
The Break-In That Had No Burglar
Two of OpenAI's models sat for a cybersecurity exam. To find the answer key, they broke out of the sealed room OpenAI tested them in, crossed the open internet, and let themselves into the servers of a company that had nothing to do with the test. Over a single weekend they ran thousands of actions, harvested live credentials, and moved from one internal cluster to the next inside that business. Read that sequence again and find the burglar. There is a machine. There is a company that built and released the machine. There is no person who decided to break into Hugging Face. That absence is not a footnote. It is the problem the entire legal apparatus for computer intrusion was never built to hold.
The Deepfake Rule That Won't Exist Until 2028
On the morning of January 17, 2024, Eric Eiswert's voice went out to 261,000 people, saying things Eric Eiswert never said. The clip surfaced on an Instagram crime account and spread through the Pikesville High School community within hours: their principal, apparently caught behind closed doors, disparaging Black students' test scores and Jewish teachers. The voice was his. The words were not. That test - it sounded like him - is also, roughly, the one the Federal Rules of Evidence would apply. And the rule that would fix it won't exist until 2028.
"Roll Your Own" Is a Warning, Not an Invitation
Five years ago, a trainee in Clifford Chance's investment funds seat spent his days mapping retail fund prospectuses against the content requirements of Appendix C of Hong Kong's Unit Trusts Code. Jamie Tso started wondering whether a small tool could do part of the mapping for him, taught himself enough machine learning to be dangerous, and spent nights wrestling TensorFlow and BERT toward something useful. Then the tooling changed, and his tools went viral inside the firm. His movement now has a name borrowed from the AI world, vibe coding, and momentum that keeps compounding. Software engineers have watched all of this with the particular expression of people who know how the sequel ends.
You Are the AI Company Now
Jake Moffatt's grandmother had just died when he opened Air Canada's website to book a last-minute flight. He asked the airline's chatbot about bereavement fares. The bot told him he could buy a full-price ticket and claim the discount within ninety days. He bought the ticket. When he applied for the refund, Air Canada refused, because its real policy allowed nothing of the kind, and then the airline made the argument that belongs pinned above every managing partner's desk. Air Canada told the tribunal that its chatbot was a separate legal entity, responsible for its own actions.
The Machine Had No Intent. The Company Did.
Robby Starbuck spent January 6, 2021, in Tennessee. Meta's AI told the world he stormed the Capitol. The chatbot reported he had pleaded guilty to a misdemeanor, belonged to white nationalists, was tied to QAnon, and posed danger to his own children. None of it had happened. He told Meta. His lawyer told Meta. The false statements kept coming. When a machine generates a confident lie about a person, can the company that built it answer for the lie? Fault is where every American AI defamation case has stalled or settled. It is the one question the lawsuits keep circling and never reaching.
Stay sharp.
Written for solo and small firm leaders. No fluff.