Supreme Court Backs FCC Fines While Robot Videos Mislead Everyone

Here is the detail that should make your jaw drop: bounty hunters and at least one rogue sheriff were using real-time cell carrier location data to track people who had absolutely no idea they were being followed. That is the backdrop to the Supreme Court case that AT&T and Verizon just lost, and it is worth keeping front of mind before feeling too sorry for two of the most profitable companies in America.

The Court ruled 8-1 this week to uphold the FCC's ability to fine the carriers a combined $104 million for selling customer location data without consent. The violations were uncovered back in 2018, the fines were issued in 2024, and the carriers have spent the intervening years trying to wriggle out of them on procedural grounds rather than arguing the underlying conduct was acceptable.

Their legal theory was creative, if a little desperate. AT&T and Verizon argued that the FCC's process for issuing financial penalties denied them their Seventh Amendment right to a jury trial. AT&T even managed to convince the Fifth Circuit Court of Appeals to buy that argument, creating a split with the Second Circuit, which ruled against Verizon. The Supreme Court stepped in to sort out the mess.

Chief Justice John Roberts, writing for the majority, was not particularly impressed by the carriers' position. His opinion pointed out that the FCC's penalty orders do not actually create a legal obligation to pay on their own. If a carrier refused to pay, the government would have to take them to court to collect, and that process would include a jury trial. The carriers always had that option. They chose to pay the fines and challenge them through the appellate system instead, which is a perfectly valid path but not one that lets you complain later about missing out on a jury.

In other words, the Court's message was roughly: you had choices, you made them, stop pretending the system was rigged.

Justice Brett Kavanaugh made an interesting observation during oral arguments, noting that the government's acknowledgment that its orders are nonbinding without court enforcement was actually a win for the carriers on the legal principle. They just could not convert that principle into avoiding the fines themselves.

The practical stakes here go beyond $104 million, which is frankly a rounding error for companies the size of AT&T and Verizon. What was really being contested was whether the FCC has a workable mechanism to investigate telecoms and propose penalties that stick. A ruling the other way would have handed carriers a blueprint for tying up any future enforcement action in constitutional challenges before a single dollar changed hands.

Advocacy groups were quick to call the decision a win for consumers, and they have a point. Phone carriers sit on some of the most sensitive location data imaginable. The idea that selling it to data brokers who then supply it to bail bondsmen and law enforcement acting outside proper channels should carry zero consequences was never a compelling one. The Supreme Court agreed, and the regulatory process for holding telecoms accountable remains intact.

The most important thing to understand about that jaw-dropping robot video in your feed is that someone, somewhere, very carefully decided what the camera would and would not show you. That is not cynicism. That is just how demonstration videos work, and robotics researchers are increasingly worried that the gap between what these clips imply and what the machines can actually do is getting dangerously wide.

The humanoid form factor makes the problem worse than it would otherwise be. When a robotic arm executes a precise movement, most people register it as a machine doing a machine thing. When a human-shaped robot does the same movement, our brains start filling in capabilities that were never demonstrated. We assume, almost automatically, that a robot that can dance could also make you a sandwich, fold your laundry, or handle whatever random task we throw at it next. That assumption is wrong, and robotics researchers say some startups are quietly counting on it when fundraising season rolls around.

Jonathan Hurst, a co-founder of Agility Robotics and a researcher at Oregon State University, put it plainly: companies do prey on that human tendency toward anthropomorphization when they need to raise money. That is not a fringe view. It is a fairly mainstream concern among people who actually build these systems.

The generalization problem is the core technical challenge nobody in the highlight reels wants to talk about. A robot that can pour a glass of wine in a controlled demo environment is genuinely impressive. A robot that can pour wine out of any bottle, into any glass, in any kitchen it has never visited before, without someone coaching it through the process, is a fundamentally different and much harder achievement. The demo proves the first thing. It implies the second. Those are not the same.

Sergey Levine, a computer scientist at UC Berkeley and co-founder of AI and robotics company Physical Intelligence, said the real benchmark for robotic capability is large-scale, quantitative evaluation in real-world environments, not curated single-take demonstrations. There is always a gap between what you can show in a demo and what a robot can actually do reliably and repeatedly. The demo is designed to close that gap visually without closing it in practice.

There is also a more basic question worth asking every time one of these videos circulates: is the robot actually autonomous, or is a human operator driving it remotely? Teleoperation, where a person controls the robot's actions in real time, can produce footage that is visually indistinguishable from full autonomy. Dipam Patel, a PhD candidate at Purdue University and a researcher with the US Army Research Lab, says that unless a company or research paper explicitly states the robot is operating without human control, healthy skepticism is warranted.

None of this means humanoid robotics is a scam. Genuine progress is happening. But there is a difference between a field advancing steadily and a field that has solved the hard parts, and right now viral video culture is doing a lot of work to blur that line.