For decades, the FBI pushed junk science on the courts, resulting in the wrongful convictions of an untold number of people. It wasn't until 2009 that it started trying to undo the damage. And even then, the FBI wasn't 100% sure it shouldn't hold onto to at least some of its favorite junk, even if it had been repeatedly shown there was very little verifiable science behind their expert witnesses' assertions.

All that seemed to matter were the convictions. Appealing a conviction is hard work -- something that takes years to do and requires the assistance of experienced lawyers. The damage has been done and the FBI's belated recognition of its contribution to the farce that is our criminal justice system isn't going to give back years of wrongfully-obtained lives.

By the FBI's own admission, "nearly every" forensic expert deployed to criminal trials gave flawed testimony that overstated the certainty of their findings. This included experts testifying about fingerprints, DNA, and hair analysis, not just those discussing complete garbage like bite-mark matching or asserting mass-produced clothing is as unique as someone's fingerprints.

The DC Appeals Court has just overturned a conviction based on faulty hair match analysis. It comes nearly fifty years after the conviction, meaning the government exchanged bad testimony for most of a person's life. The opening of the decision [PDF] lays out the facts concisely.

Almost fifty years ago, appellant Dennis Butler was convicted of murder. At his trial, an FBI forensic expert testified that hairs found on the victim were microscopically identical to Butler’s hair. The government recently acknowledged, though, that hair evidence of the kind introduced against Butler was false and exceeded the limits of science, and that the prosecution knew or should have known as much at the time of his trial.

Butler has spent decades in prison because the FBI spent decades saying having a microscope was the same thing as being a scientist. The court says it's not even close… and it's a statement made using the FBI's own findings.

Hair microscopy called for forensic examiners to conduct side-by-side, microscopic comparisons of hair samples in an effort to ascertain whether hairs from a crime scene matched hairs from a suspect. The government used ostensible matches at trial as scientific evidence linking defendants to crimes.

There was, however, a significant problem with that field of analysis: science had not validated its foundational premises. Existing studies failed to support a trained examiner’s ability to identify a “match” based on any objective system of visual hair comparison or to validly estimate the frequency of hair characteristics (and therefore of matches) in the general population.

This lack of scientific confirmation didn't stop the FBI from using hair matches in court, even after it knew the "science" behind it was lacking. It wasn't until the National Academy of Sciences called the FBI out in its report on hair analysis that the FBI began reining in its expert witnesses. Following this public fisking by actual scientists, the government reviewed thousands of convictions obtained with hair analysis prior to the year 2000. Butler's was one of them.

So, the question is: did the prosecution rely on the FBI's faulty testimony to secure this conviction? The answer is "yes." The government's evidence (beyond the supposed hair match) were two unreliable witnesses whose stories changed frequently, some paint chips (from the recently-painted apartment Butler had visited while it was being painted), and some eyewitnesses who put Butler near the crime scene (with the prosecution dodging the fact that Butler was often in the area where the murder took place for non-murderous reasons).

Here's the prosecutor during closing arguments, discussing the expert witness' testimony.

He said when he compared the hairs that were found on the victim’s clothing with the defendant’s hairs that were taken by [the detective] from him at the infirmary, when he compared those two, what were they? They were the same in every microscopic detail, the same. I said, how often . . . does it happen? You can’t be positive, yes, but how often does it happen that two people’s hair, two different people, are so similar and so alike that you would be unable to tell? Out of 10,000 examinations, he said he recalls it happening approximately four times.

And again:

You have the FBI report saying that this man’s hair compared with the hairs found on the body of the dead man. They are the same in every microscopic characteristic—every one. You heard the sixteen possible combinations, lack thereof, etc. Every one matched.

This influenced the jurors, the Appeals Court says. The jurors could have found the other evidence circumstantial, at best. But the repeated (incorrect) assertion that hair match analysis could positively identify Butler as the murderer most likely swayed the jury, resulting in a conviction that resulted in Butler being behind bars since late 1970.

We found it satisfied in Ausby, and we do likewise here. And we thus conclude that the government’s presentation against Butler of evidence that it knew (or should have known) was false denied him a fair trial.

For the foregoing reasons, we reverse the judgment of the district court and remand with instructions to grant Butler’s motion to vacate his sentence pursuant to 28 U.S.C. § 2255.

The FBI convinced itself it could craft new forensic techniques out of thin air and deploy them without making sure they actually had some scientific basis. This is the end result of this hubris -- acts that took decades to correct and even longer to result in actual justice, rather than unearned convictions. Some of the people screwed by the FBI's junk science are already dead. Others, like Butler, may finally see their lives given back to them -- but with hardly any life left to live.

Techdirt has been writing about right-to-repair laws -- or, rather, their absence -- for many years now. A recent right-to-repair post concerned ventilators, pretty much the last hope for critically-ill patients suffering the effects of the new coronavirus. This underlines the fact that being able to repair equipment you have bought is not an abstract issue, but is literally a matter of life or death in some cases. Despite that, in Australia the fight to obtain a right to repair is still in its early stages:

The 'right to repair' movement has finally bent the ear of Australia's competition and consumer watchdog, the ACCC, in its pleas to be able to fix their own farm equipment.

An ACCC inquiry will examine whether international tractor manufacturers are failing Australian farmers who want access to software tools and parts to repair their own machinery.

As that news item on the Australian Broadcasting Corporation's site indicates, the impetus for a right-to-repair law comes from the agricultural sector. Five years ago, Techdirt wrote about a similar case in the US, which involved tractors from John Deere. The ACCC released a discussion paper on the topic at the end of February. It seeks feedback on what it terms four "concerns", specifically that:

1. access to independent agricultural machinery repairs is limited

2. farmers may lack recourse in the event of a problem with their machinery

3. agreements between manufacturers and dealers may limit access to repairs

4. data ownership and management may raise privacy and competition issues.

The last of these is particularly interesting. It reflects the increasing sophistication of the once-humble tractor, which now involves both software and data. The ABC story explains:

In an era of water scarcity and a swelling global population, machinery makers have poured millions of dollars into developing software that allows farmers to precisely plot their sprawling properties, gauging how much seed, water, fertiliser, and pesticide is needed for maximum crop yields for each field.

It is clearly vital for farmers to retain control over their own data, while equipment manufacturers see this as a resource they can control and exploit -- for example, by aggregating data from many farms and selling access to it. People owning agricultural equipment -- or anyone else -- have until Sunday, 31 May 2020 to make their submissions to the Commission. It will be a while before the ACCC reports on what it finds, and after that the battle to enshrine a right to repair in Australian law will probably take years. But at least the process has begun, which could give a useful impetus to other efforts around the world to bring in similar, much-needed legislation.

Follow me @glynmoody on Twitter, Diaspora, or Mastodon.

It has been nearly seven years since 18-year-old Graham Dyer died due to injuries he sustained while riding in the back of Mesquite (TX) Police car. Dyer, all of 5'4" and 110 pounds, was picked up by Mesquite officers while experiencing a bad acid trip. Dyer had no idea what was happening to him or where he was. As the officers transported him to jail, he thrashed around in the back of the patrol car, ultimately slamming his head into the seat, window, and metal bars forty-six times.

The officers never bothered to restrain him. But they did stop the car to head into the back seat to tase him, including one prolonged burst (eight seconds) delivered directly to Dyer's testicles. Dyer's parents had no idea what had happened to their son, who was picked up in good health (bad trip notwithstanding) but was dead less than 24 hours later. The Mesquite PD refused to release any records of this arrest to Dyer's parents. The Dyers finally obtained some records, but from the FBI, which had been called in to investigate the in-custody death.

Armed with these records, the Dyers sued the officers and paramedics who handled the arrest and the less-than-adequate care their son received. The paramedics have been dismissed from the lawsuit, but most of the allegations against the officers survive. Two of the three officers involved are still facing the Dyers' lawsuit. And the Fifth Circuit Appeals Court has just thrown the third officer back into the mix, stripping the qualified immunity the lower court awarded him.

The Appeals Court says the lower court made the right call for two of the officers, who are facing deliberate indifference claims for apparently not caring at all the arrestee they were delivering to jail was possibly seriously injured. From the decision [PDF]:

The district court correctly found a genuine dispute concerning whether Gafford and Heidelburg were deliberately indifferent to the serious medical needs of a detainee in their custody. A reasonable trier of fact could find that those Officers were aware that Graham, in the grip of a drug-induced psychosis, struck his head violently against the interior of Heidelburg’s patrol car over 40 times en route to jail and thereby sustained severe head trauma. Both Officers told Graham to stop hitting his head and Heidelburg even pulled his patrol car over in an effort to stop him. Gafford acknowledged that, during his encounter with Graham, he knew “[t]here could be some inherent dangers” associated with head trauma; Heidelburg testified that what Graham was doing “certainly could” cause a head injury. Yet the Officers sought no medical care for Graham when they arrived at the jail. Nor did they alert jail officers (who had no way of knowing what had happened en route to the jail) of the possibility that Graham had seriously injured himself. The record instead reflects that the jail sergeant was “[i]nformed by [the] transport officers [that] Dyer had been medically cleared at the scene.”

As the court noted there, "no officer" informed jail staff about Dyer's possible injuries. That includes Officer Zachary Scott, who was also aware of Dyer's unrestrained thrashing in the back of the patrol car, but similarly chose to say nothing about it.

The district court found the Dyers “failed to present evidence that [Scott was] aware of facts indicating a risk of injury and inferred a risk of injury to Graham.” True, Scott’s affidavit stated he “never observed anything or any action by anyone which might cause a head injury on the part of [Graham],” nor did he “observe[] anything to indicate [Graham] might have any serious injury.” But his deposition testimony was quite different. Scott testified he assumed Heidelburg pulled the patrol car over because Graham “was banging his head.” Scott also testified he tried “to prevent [Graham] from banging his head on the back of the car.” Lastly, Scott stated he did not tell the jail sergeant about Graham slamming his head, nor did he recall hearing anyone else report it.

None of the officers seemingly cared if Dyer lived or died or even got a second look from jail medical staff. Dyer died and now all three officers will face deliberate indifference claims from Dyer's parents.

[I]n this case a reasonable jury could find that (1) Graham violently bashed his head against the interior of Officer Heidelburg’s patrol car over 40 times while en route to jail; (2) Officers Heidelburg, Gafford, and Scott were fully aware of Graham’s actions and of their serious danger; (3) the Officers sought no medical attention for Graham; and (4) upon arriving at jail, the Officers failed to inform jail officials what Graham had done to himself, telling them only that Graham had been “medically cleared” at the scene. From this evidence, a reasonable jury could conclude that the Officers “were either aware or should have been aware, because it was so obvious, of an unjustifiably high risk to [Graham’s] health,” did nothing to seek medical attention, and even misstated the severity of Graham’s condition to those who could have sought help.

Let's hope the jury is reasonable, because everything about these officers' actions isn't. The officers chose to handle a small teen's horrific drug reaction by subjecting him to further horrors, including several tasings. The audio from the in-car camera caught one officer telling the traumatized Dyer, "Motherfucker, I'm going to kill you." And so he did. Maybe not immediately, but that was the end result of these officers' actions -- actions that included choosing not to inform anyone about the multiple head injuries Dyer suffered while in their "care."

We keep trying to explain to people that privacy is always about trade-offs, and arguing for privacy laws that protect "privacy" as if it's a constant thing, will run into trouble. Most of that trouble is in the form of locking in big companies, but sometimes, the trouble is in showing you why understanding trade-offs matters so much.

France has been among the most vocal critics of "big internet companies" and demanding various regulatory pressures be used to punish them. Last year it fined Google $57 million for breaching privacy laws, and appears to be angling for even larger fines.

So it's difficult not to burst out in laughter after finding out that the French government is really, really mad that Google and Apple are protecting people's privacy, when suddenly the French government wants to use those companies to engage in contact tracing. Indeed, it's literally demanding both companies ease their privacy protections to help France track people who might have COVID-19.

France has become the first country to call publicly for Apple and Google to weaken privacy protections around digital contact tracing, after its government admitted that its current plans would not work without changes to smartphone operating systems.

The issue is that, as you likely recall, earlier this month, Google and Apple collaborated on an API system to enable some form of contact tracing in various apps, but that is (a) voluntary, (b) privacy retaining (via regularly rotating identifiers), and (c) limits how much information would be sent to the government. And apparently that's getting in the way of France's more aggressive tracking plans:

France wants to deploy its app by 11 May, without using the special measures Apple and Google have put in place, which are targeted for release in mid-May. That means the country will be forced to use the more limited features already built into iOS, unless Apple changes its policies and allows for far more invasive use of the Bluetooth radio at the heart of its devices.

It's quite incredible to see this play out in practice, in France of all places. Again, the French government has been among the most vocal and aggressive in attacking Google and insisting that its privacy practices are terrible... but as soon as those privacy efforts get in the way of the French government spying on people's whereabouts, its suddenly mad at these companies for doing too much to protect privacy? Maybe Google should see if the French government wants to pay back the fines it levied before it'll take the government's requests seriously.

"See something, say something" but for the coronavirus. How could it possibly go wrong?

Everyone put your hands down. Of course we know how it could go wrong. Opening up the lines to callers never works. The DHS knows this, even if it chooses to ignore this. Years of "see something, say something" did nothing more than fill its data stores with reports from curtain twitchers about things their brown neighbors were doing and do-gooding randos calling in everyone they saw walking around with a camera.

The home to the most "every Muslim is a suspect" city in the US should have known better than to open up a snitch line to the public. But officials like Governor Andrew Cuomo have already shown they're unable to fathom the concept of unintended consequences. Cuomo issued a mandate for mask-wearing in public while allowing an anti-mask law to run concurrently, inviting state law enforcement officers to engage in very selective enforcement.

Without a doubt, the state has a COVID problem. But doing dumb shit isn't going to fix it. At best, it's just going to tie up city resources.

Enter NYC Mayor Bill De Blasio. A city in the deep throes of a pandemic needs solid leadership. This isn't it. This is only something that seems sort of useful until someone gives it a few seconds of thought.

On April 18, Mayor De Blasio took decisive action in an effort to curb the spread of the coronavirus:

If you see someone failing to practice social distancing, rat them out, Mayor Bill de Blasio urged Saturday as part of a new city effort to corral the coronavirus.

“We still know there’s some people who need to get the message. And that means sometimes making sure the enforcement is there to educate people and make clear we’ve got to have social distancing,” Hizzoner said in a video posted to Twitter

To do that, is “simple,” he explained.

Snap a photo of an offending person or crowd, set the location on the image, and text it to 311-692.

I don't know what he expected, but anyone who's had any internet experience knows what the public will do when it's given access to a government hotline (or an internet poll, etc.).

Photos of extended middle fingers, the mayor dropping the Staten Island groundhog and news coverage of him going to the gym have all been texted to a special tip line that de Blasio announced Saturday, according to screenshots posted on Twitter.

Believe it or not, that's the better stuff. The government wanted the public to respond and respond the public did. Emphatically and obscenely.

An NYPD source said that “dick pic” photos of real penises have also been texted to 311, and a caller phoned in a tip that de Blasio was seen performing oral sex on someone “in an alleyway behind a 7-11” early Sunday.

Feeling any safer, New York? If you ask people to become informants, you're going to have to sift through a lot of bullshit. If you ask people to report their neighbors for being in the proximity of other people, you're going to get Hitler references and dick pics. And if you haven't prepared for this inevitable eventuality, you're going to have to cut off your access to the very small amount of genuine tips to clean out all the garbage you didn't realized you'd asked for.

City Hall confirmed Tuesday that the 311 text line was shuttered to clear out a backlog of crank complaints on Sunday.

The solution isn't a new tip line that's going to quickly turn into an IRL 4Chan. The solution is to let existing systems handle the influx and hope for the best. Yes, too many people still use 911 and dispatch lines like all-purpose complaint boxes but asking New Yorkers to snitch on each other for being social was never going to work out well.

People are already feeling the encroachment of the government as the pandemic wears on. Asking residents to be part of the perceived problem is always going to result in little-a anarchy. Government officials never seem to understand most Americans think being a snitch is pretty much un-American. This is something we expect authoritarian nations to do: turn citizens against each other. We don't care for it much in the Land of the Free. There will always be outliers willing to sell the person across the street out to the fuzz, but most Americans would rather let their fingers (or other appendages) fly when the government asks them to be unpaid informants.

Digitization has conquered our lives, both personally and professionally — transporting almost all data into the vast world of internet, networks, clouds, servers, and more. And while this has made convenience and speed in terms of access, threats and risks are close at hand too. The Complete Cybersecurity Training Bundle will give you 90+ hours of content on basic to advanced cybersecurity skills covering CompTIA, CISA, CISSP, Cloud Computing, COBIT 5, and PMI-ACP. It's on sale for $40.

Note: The Techdirt Deals Store is powered and curated by StackCommerce. A portion of all sales from Techdirt Deals helps support Techdirt. The products featured do not reflect endorsements by our editorial team.

Richard Epstein is a very famous law professor, known for his "libertarian" take on the world. Lots of people who know him insist he's a brilliant legal mind... who seems to think that his brilliance in that area allows him to be brilliant in fields where he has no experience at all. For years, I've followed him being just ridiculously wrong when it comes to internet law and (even more so) on any issue related to copyright or patents, which he views as identical to tangible property. He has long refused to even consider that he might be wrong about that. Still, it was pretty shocking last month to see him jump into the deep end of the debate over the seriousness of COVID-19 by writing a piece claiming that he expected US deaths to top out at 500 tops. This was on March 16th, at which time California and Washington were already shutting down and it was blatantly clear many more people would die. Still what he initially wrote was:

[I]t seems more probable than not that the total number of cases world-wide will peak out at well under 1 million, with the total number of deaths at under 50,000 (up about eightfold). In the United States, if the total death toll increases at about the same rate, the current 67 deaths should translate into about 500 deaths at the end.

Remember, Epstein is a law professor, not an epidemiologist, not a statistician, and not an economist (though he often acts as if he is one). That pronouncement, which bolstered the claims coming from the White House and Trump's supporters at the time that COVID-19 was still no big deal, quickly made its way to the White House, because clueless powerful people love an echo chamber.

As the death toll quickly passed 500, Epstein hastily revised his article to say he actually meant 5,000 US deaths, and that he regretted what he called "a simple stupid error" in his calculations. Epstein was roundly mocked for all of this, and continues to appear painfully unwilling to recognize his own ignorance in this issue, as was evidenced in this truly incredible interview Epstein gave with the New Yorker's Isaac Chotiner in which he tries to defend his indefensible argument, and comes off like a total condescending asshole who assumes that because he's a law professor he somehow knows more about epidemiology than a journalist who he insists must know nothing at all.

I know, but these are scientific issues here.

You know nothing about the subject but are so confident that you’re going to say that I’m a crackpot.

No. Richard—

That’s what you’re saying, isn’t it? That’s what you’re saying?

I’m not saying anything of the sort.

Admit to it. You’re saying I’m a crackpot.

I’m not saying anything of the—

Well, what am I then? I’m an amateur? You’re the great scholar on this?

No, no. I’m not a great scholar on this.

Tell me what you think about the quality of the work!

O.K. I’m going to tell you. I think the fact that I am not a great scholar on this and I’m able to find these flaws or these holes in what you wrote is a sign that maybe you should’ve thought harder before writing it.

What it shows is that you are a complete intellectual amateur. Period.

O.K. Can I ask you one more question?

You just don’t know anything about anything. You’re a journalist. Would you like to compare your résumé to mine?

No, actually, I would not.

Then good. Then maybe what you want to do is to say, “Gee, I’m not quite sure that this is right. I’m going to check with somebody else.”

Of course, Chotiner did check with many other people and added multiple fact checks to his interview, quoting actual experts who respond with things like saying that what Epstein is saying is "completely inaccurate."

Incredibly, earlier this week, Epstein tried again, with a new piece complaining that governors are overreacting with their lockdown decrees. Incredibly, he uses the evidence that these lockdowns are working and bending the curve to decrease the number of deaths as evidence that we should end the lockdowns, and that magically businesses will be able to handle the rest.

The question is why Cuomo thinks that doubling down on government restrictions is justified by the science and data. His own daily report of April 17 indicates that the rate of new infections is down and that the number of hospital discharges in the state now far exceeds the number of new admissions. Further data prepared by the New York Times reveals that the rate of infection is now slowing down rapidly throughout the United States. It also shows that new cases peaked at about 35,000 on April 3, with an erratic decline since that time.

That's because of the lockdown. If we lift the lockdown too soon, those infection and death rates go back up.

Still, when the article came out, I went to go look for the original Epstein piece claiming only 500 deaths, and was surprised to find that the link now contains this:

[Correction & Addendum as of March 24, 2020:

My original erroneous estimate of 5,000 dead in the US is a number ten times smaller than I intended to state, and it too could prove somewhat optimistic. But any possible error rate in this revised projection should be kept in perspective. The current U.S. death toll stands at 592 as of noon on March 24, 2020, out of about 47,000 cases.

So... uh... what? Everything about that appears to be gas lighting. First off, the original estimate was 500 deaths, not 5,000. And he claims that was due to a "simple stupid error." But now that we're rapidly approaching 50,000 deaths, he's claiming that it was the 5,000 that was the "original erroneous estimates" and that he "intended to state" 50,000? Really?

I was confused and really wondered what happened to that original 500 estimate, and thankfully John Macke figured it out making use of the Internet Archive's Wayback Machine, and it appears that someone did some tricky shit to try to hide the original estimate -- whether it was Epstein directly or someone else at the Hoover Institution:

Of course, now that those revisions have been caught and started to go viral, with little fanfare (and it appears to have happened while I was writing this post, someone quietly added a new correction to the old corrected correction (got that?):

[APRIL 21, 2020 EDITOR'S NOTE: DUE TO AN EDITING ERROR, THE CORRECTION & ADDENDUM ABOVE IS INACCURATE. IT SHOULD STATE:

That estimate is ten times greater than the 500 number I erroneously put in the initial draft of the essay, and it, too, could prove somewhat optimistic. But any possible error rate in this revised projection should be kept in perspective. The current U.S. death toll stands at 592 as of noon on March 24, 2020, out of about 47,000 cases. So my adjusted figure, however tweaked, remains both far lower, and I believe far more accurate, than the common claim that there could be a million dead in the U.S. from well over 150 million coronavirus cases before the epidemic runs its course.

WE REGRET THE ERROR AND ANY CONFUSION IT HAS CAUSED]

But again, that "editing error" was only introduce weeks later after the "correct" error message was there. What would make them go back and change an error message that was shown for two weeks and make it false? Is it possible that an editor somehow went in and changed around the correction to take it from accurate to false? Perhaps, but it still seems... quite odd.

Of course, the bigger question is why should anyone ever take Epstein's ideas on all this seriously, given his dreadful track record, and lack of expertise in this particular field?

We've noted repeatedly that despite a lot of talk from U.S. leaders and regulators about the "digital divide," the United States doesn't actually know where broadband is available. Historically the FCC has simply trusted major ISPs -- with a vested interest in downplaying coverage and competition gaps -- to tell the truth. The FCC's methodology has also long been flawed, considering an entire area to be connected if just one home in a census tract has service. The results are ugly: the FCC's $350 million broadband availability map all but hallucinates broadband availability and speed (try it yourself).

As pressure mounts on the agency to finally improve its broadband mapping, the scope of the problem continues to come into focus. Like this week, when AT&T was forced to acknowledge that the company provided the FCC with inaccurate broadband availability data across 20 states, impacting some 3,600 census blocks:

"AT&T disclosed the error to the FCC in a filing a week ago. The filing provides "a list of census blocks AT&T previously reported as having broadband deployment at speeds of at least 25Mbps downstream/3 Mbps upstream that AT&T has removed from its Form 477 reports." The 78-page list includes nearly 3,600 blocks."

You'll recall that last year, Ajit Pai tried to claim that his "deregulatory agenda" (read: gutting oversight of an uncompetitive and hugely unpopular business sector) resulted in some amazing broadband expansion. Only later was it revealed that much of this growth either was triggered by things Pai's FCC had nothing to do with (like fiber build-out requirements affixed to AT&T's 2015 merger with DirecTV by the previous FCC), or a broadband mapping blunder by a small provider by the name of BarrierFree, which overstated its footprint to the FCC by a cool 1.5 million locations.

AT&T insists this latest error was caused by a "software bug," and while relatively small in the scope of AT&T's overall service area, consumer groups are a little curious how it could have gone unnoticed for the better part of two years:

"Aside from one even bigger error by an ISP called BarrierFree last year, Turner said he hasn't "seen any other ISP reporting error like this before" and that "it is curious that the [AT&T] error may have gone unnoticed for 2-plus years."..."While relatively small errors like this don't end up changing conclusions about national trends, it certainly can impact the FCC decisions about where to spend—and where to not spend—scarce subsidy funds," Turner said. "AT&T should be quite a bit more forthcoming about the exact nature of this error and how it discovered it, so that other ISPs can be sure they're not making similar errors."

While there's no evidence of intentional under-reporting by AT&T, the timing is curious all the same.

After several decades of complaints, pressure has mounted on the FCC and Congress to actually do something as states vie for additional deployment subsidies. That culminated in the recent passage of the Broadband Deployment Accuracy and Technological Availability (DATA) Act, which mandates the FCC to use more accurate geolocation and crowdsourced data to create more accurate maps and actually verify where broadband's available before doling out billions in subsidies or issuing policy (fancy that!).

It will take years to complete, the FCC has warned they can't afford to finish it without more funding, and the industry, which has spent years lobbying against mapping improvements for obvious reasons, could still find ways to either scuttle the effort or make access to the data difficult. Still, baby steps and all that. There are at least indications that the "what US broadband competition problem?" telecom policy set finally realizes this is a problem that needs fixing, even if truly better broadband maps are still several years away.

All the way back in 2014, Twitter sued the DOJ over its National Security Letter reporting restrictions. NSLs are the FBI's weapon of choice in all sorts of investigations. And they almost exclusively come packaged with lifetime bans on discussing them publicly or disclosing the government's request for info to NSL targets.

Things changed a little with the passage of the USA Freedom Act and a couple of related court decisions. The DOJ is now required to periodically review NSLs to see if the ongoing silence is justified. The Act also finally provided a way for companies to challenge gag orders, which has resulted in a somewhat steady stream of published NSLs.

What's still forbidden is publishing an actual count of NSLs a company has received. Supposedly the security of the nation would be threatened if Twitter said it had received 118 NSLs last year, rather than "0-499." The reforms in the USA Freedom Act didn't change that aspect of NSL reporting and the government still argues any accurate reporting would allow the terrorists to win… or somehow avoid being targeted by an NSL.

Twitter argued the publication of an accurate number was protected speech. The government, of course, argued the opposite. The federal judge handling the case ruled that accurate reporting wasn't protected speech back in 2016, but did say Twitter could move forward with its challenge of the classification of this data.

Roughly a year later, the court changed its mind. The government's motion to dismiss was denied by the court, which said it needed to come up with better arguments if it wanted to escape Twitter's lawsuit. The court pointed out that denying Twitter the right to accurately report NSLs was a content-based restriction that couldn't be justified by the government's bare bones assertions about national security.

Nearly three years later, we're back to where we were four years ago. The court has dismissed Twitter's lawsuit, denying its attempt to escape the "banding" restrictions that limit the transparency it can provide to its users. (via Politico)

The decision [PDF] -- which ends nearly six years of litigation -- says the court believes the things the government says about detailed NSL reporting. Since these declarations tend to be delivered in ex parte hearings and/or under seal, we have to believe them, too. Actual numbers are more dangerous than vague numbers.

The declarations explain the gravity of the risks inherent in disclosure of the information that the Government has prohibited Twitter from stating in its Draft Transparency Report, including a sufficiently specific explanation of the reasons disclosure of mere aggregate numbers, even years after the relevant time period in the Draft Transparency Report, could be expected to give rise to grave or imminent harm to the national security. The Court finds that the declarations contain sufficient factual detail to justify the Government’s classification of the aggregate information in Twitter’s 2014 Draft Transparency Report on the grounds that the information would be likely to lead to grave or imminent harm to the national security, and that no more narrow tailoring of the restrictions can be made.

And that's it. The restrictions stay in place and recipients of NSLs will only be able to deliver government-approved information about them. The good news is there's a bit of a loophole -- one the court discusses in a footnote. The DOJ may want to restrict almost all NSL reporting, but the court isn't convinced companies can be limited to using the DOJ-approved "bands" only.

The [complaint] alleges a facial constitutional challenge to FISA’s secrecy provisions to the extent they categorically prohibit the reporting of aggregate data. The Court does not find that they do so restrict the aggregate data at issue here. The Government has, in part, argued that FISA’s statutory nondisclosure provisions, applicable to the existence and contents of individual orders, logically prohibit reporting of aggregate data about the number of such orders. The Court has never found the Government’s logic persuasive on this point. The requirement not to disclose a particular order is completely distinct from disclosing the aggregate number of orders.

This seems to say companies can accurately report the total number of NSLs they've received, rather than using the far more vague 0-499, etc. reporting they've been limited to. It's not a lot but it's an improvement.