Here's a detail that should stop you mid-scroll: federal agents tracked a man down to his hotel room — hours after he landed from a vacation in Finland — because of a strongly worded email he sent to a government official.
David Streever, a Rochester resident, sent an angry message to ICE's then-acting director Todd Lyons after federal agents killed two US citizens in Minneapolis. The email was blistering, comparing Lyons to a Nazi war criminal and predicting his eventual disgrace. It was the kind of thing people type in furious moments and sometimes actually send. Streever sent it.
What happened next is the part that matters. While Streever was abroad with his daughter, DHS agents showed up at his home and handed his wife a formal warning notice. Then, almost the moment he stepped back on US soil, another agent located him at his Manhattan hotel. The notice informed Streever that his email "MAY BE IN VIOLATION OF FEDERAL LAW." No arrest. No charges. Just a knock on the door and a piece of paper designed to make him wonder what comes next.
Streever has since filed a federal lawsuit, and his attorney Adam Steinbaugh from the Foundation for Individual Rights and Expression put the legal issue plainly: the First Amendment doesn't only protect you from being arrested for your speech. It also protects you from the government showing up to intimidate you into silence. The visit to the hotel, the warning notice, the agents talking to his wife — none of that requires handcuffs to be a rights violation.
This isn't a one-off. ICE's Office of Professional Responsibility has opened more than 100 investigations into alleged "doxing" and threats involving ICE agents over the past year, according to reporting from Wired. The catch: a lot of what DHS is calling doxing involves sharing information about federal agents that is already publicly available. Criticizing an official by name, or circulating their publicly listed work information, is a long-standing tradition of political speech in this country.
The pattern here is what makes this story genuinely alarming. No criminal charges have been filed in these cases. But that almost seems beside the point. When agents show up at your home and track you across state lines over a heated email, the message gets delivered without a single prosecution. The chilling effect is the punishment.
DHS reportedly works with web platforms to identify critics, a practice that raises its own set of questions about how private user data flows to federal law enforcement. The government doesn't need to win in court to win. It just needs people to think twice before hitting send.
In a political environment where the immigration crackdown has been a centerpiece of the administration's identity, the stakes of criticizing that crackdown are rising. That's not a coincidence. It's a strategy.
AI
Ivy League Professor Catches AI Cheaters With In-Person Exam
Forty students in a notoriously difficult Brown University economics course scored a perfect 100 on their midterm. The historical average for that same exam sits somewhere between 65 and 80 percent. The professor made this particular version harder than usual. Something did not add up.
Roberto Serrano, a blind economics professor at Brown, had made an emotional decision after a campus shooting in December 2025 to shift his spring ECON 1170 course to take-home exams. The class, which typically draws fewer than 30 students, suddenly attracted 86 enrollees. When the midterm results came back with an average score of 96 out of 100, Serrano didn't celebrate. He got suspicious.
The answers felt strange. Technically correct in many cases, but written in what Serrano described as a "very convoluted style" that didn't match how his students typically expressed themselves. He and his graduate students ran the exam questions through ChatGPT and got back answers that looked remarkably familiar.
So Serrano made the final exam in-person. He told his class exactly what he was doing and why, framing it as a chance to prove him wrong. They did not prove him wrong. Scores collapsed. The gap between the take-home midterm performance and the in-person final was dramatic enough to make the scale of the cheating undeniable. A class that had collectively scored near-perfect on one exam suddenly looked like a completely different group of students when ChatGPT wasn't in the room.
Serrano has been vocal about the episode, speaking to El Pais and Inside Higher Ed in the past week. His frustration isn't just about academic integrity in the abstract. His core argument is sharper than that: using AI to complete work you're supposed to be learning from doesn't just break the rules, it breaks the student. You cannot shortcut your way to competence. The exam is not the point. The thinking is the point.
This story lands at an interesting moment. A Princeton survey recently found that nearly 30 percent of students admitted to using AI to cheat on at least one exam or assignment. Brown's situation puts flesh on that statistic. It shows what that cheating actually looks like at scale inside a single classroom — a mass enrollment spike when take-home exams are announced, suspiciously perfect scores, and then a very public unraveling.
The harder question underneath all of this is what universities are supposed to do about it. AI detection tools are unreliable and have falsely flagged legitimate student work. In-person exams are one answer, but they're not always practical. And motivated students who want to cheat will keep finding new angles.
Serrano's approach — transparency, documentation, and refusing to quietly let it go — is at least an honest response to a problem that higher education has mostly been tiptoeing around. Whether it changes anything beyond his classroom remains to be seen.
Source: Ars Technica
Roberto Serrano, a blind economics professor at Brown, had made an emotional decision after a campus shooting in December 2025 to shift his spring ECON 1170 course to take-home exams. The class, which typically draws fewer than 30 students, suddenly attracted 86 enrollees. When the midterm results came back with an average score of 96 out of 100, Serrano didn't celebrate. He got suspicious.
The answers felt strange. Technically correct in many cases, but written in what Serrano described as a "very convoluted style" that didn't match how his students typically expressed themselves. He and his graduate students ran the exam questions through ChatGPT and got back answers that looked remarkably familiar.
So Serrano made the final exam in-person. He told his class exactly what he was doing and why, framing it as a chance to prove him wrong. They did not prove him wrong. Scores collapsed. The gap between the take-home midterm performance and the in-person final was dramatic enough to make the scale of the cheating undeniable. A class that had collectively scored near-perfect on one exam suddenly looked like a completely different group of students when ChatGPT wasn't in the room.
Serrano has been vocal about the episode, speaking to El Pais and Inside Higher Ed in the past week. His frustration isn't just about academic integrity in the abstract. His core argument is sharper than that: using AI to complete work you're supposed to be learning from doesn't just break the rules, it breaks the student. You cannot shortcut your way to competence. The exam is not the point. The thinking is the point.
This story lands at an interesting moment. A Princeton survey recently found that nearly 30 percent of students admitted to using AI to cheat on at least one exam or assignment. Brown's situation puts flesh on that statistic. It shows what that cheating actually looks like at scale inside a single classroom — a mass enrollment spike when take-home exams are announced, suspiciously perfect scores, and then a very public unraveling.
The harder question underneath all of this is what universities are supposed to do about it. AI detection tools are unreliable and have falsely flagged legitimate student work. In-person exams are one answer, but they're not always practical. And motivated students who want to cheat will keep finding new angles.
Serrano's approach — transparency, documentation, and refusing to quietly let it go — is at least an honest response to a problem that higher education has mostly been tiptoeing around. Whether it changes anything beyond his classroom remains to be seen.
Enjoyed this?
Get stories like this delivered every Tuesday — free.