A couple of years back, we discussed a win in the courts for the U.S. Dairy Export Council against both French and Swiss consortiums that had attempted to trademark the word “Gruyere” in America. Both of those groups were jointly attempting to get a PDO/PGI trademark for the term, arguing that gruyere cheese should only be labeled such if it came from the region around Gruyeres, a Swiss town somewhat near the French border. This style of trademark is the sort used by consortiums such as that for champagne, which has successfully argued for protected status on products based on geographic origin. In this case, both the USPTO and the U.S. courts rightly concluded that these consortiums could not unring this bell. Americans have been eating gruyere cheeses for years and years, with the vast majority of them likely not even knowing that the town of Gruyeres existed, never mind being confused into thinking that the cheese they bought at the store must have come from that region.
While both groups claimed they planned to appeal the ruling, that seems less likely now. Recently, the Swiss government declined a motion to continue defending the gruyere name around the world, instead softening its language around what the government would be doing about it. That motion was in reaction to the United States’ court decision, meant to force the Swiss government to include such protections in future trade deals.
On Monday, the Swiss Senate tacitly rejected a motion aimed at better defending PDO and PGI (protected geographical indication) products like Gruyère cheese. It did, however, say it was prepared to raise awareness of these labels.
The House of Representatives had approved the motion last year, after Gruyère cheese’s PDO was definitively abandoned by American retailers.
The text of the motion was changed to remove the requirement that the government enforce such protections. The Swiss Senate correctly decided that this would hamstring the government’s ability to engage in and adopt new trade agreements out of deference to the cheese-people should a partner country not agree to those protections.
So instead of an enforcement mechanism, the government is instead supposed to take educational and advocacy measures instead.
Such a measure would restrict Switzerland’s room for manoeuvre in terms of foreign trade, and could even prevent the conclusion of new agreements, declared Economy Minister Guy Parmelin.
However, the Senate has proposed an alternative to the motion, with very few constraints. It encourages the government to make a greater commitment to respecting the recognition of products of Swiss origin. Awareness-raising could, for example, be carried out as part of trade agreement negotiations. This adapted motion was tacitly approved by the Senate and now goes to the House of Representatives for approval.
The updated motion still needs to be passed by the Swiss House, but it’s obviously a better plan than the original motion. Sure, making sure people understand the history of this cheese as part of Swiss culture can be a good thing. Educating the consumer about the origins of that style of cheese may also drive foreign consumers to want to buy gruyere cheeses from its historical birthplace, too.
But what it won’t do is lock up the name of what has become a style of cheese in foreign lands like America that consumers don’t currently associate with that birthplace.
In this week’s roundup of the latest news in online speech, content moderation and internet regulation, Mike is joined by guest host Professor Kate Klonick, who has studied and written about trust & safety for many years and is currently studying the DSA & DMA in the EU as a Fulbright Scholar. They cover:
This episode is brought to you with financial support from the Future of Online Trust & Safety Fund, and by our sponsor TaskUs, a leading company in the T&S field which provides a range of platform integrity and digital safety solutions. In our Bonus Chat at the end of the episode, Marlyn Savio, a psychologist and research manager at TaskUs, talks to Mike about a recent study they released regarding frontline moderators and their perceptions and experiences dealing with severe content.
Going back many, many years, we’ve written about how the public narrative that the large social media networks engage in “anti-conservative bias” in their content moderation policies is bullshit. Because it is. And now we have yet another scientific study to prove this.
The first time we covered it was in response to a ridiculous “study” from MAGA darling and “unremarkable racist” Richard Hanania. He released a ridiculously embarrassing study claiming to “prove” anti-conservative bias in Twitter suspensions. It was based on a self-collected set of just 22 examples of accounts suspended from Twitter, in which Hanania noted that 21 of the 22 accounts were “Trump supporters.” What he left out of his analysis was that a bunch of those “Trump supporters” were… out and out neo-Nazis, including (I’m not joking) the American Nazi Party account.
Since then, many other actual studies have called bullshit on the claims of anti-conservative bias in moderation. Indeed, the evidence has suggested that both Twitter and Facebook even adjusted the rules to allow for even greater levels of rules violations for MAGA supporters, just to avoid the appearance of anti-conservative bias. That is, their bias was actually pro-MAGA in that they loosened the rules for Trump-supporting accounts, allowing them to break the rules more frequently.
This is what people mean when they talk about “working the refs.” So much of the whining and complaining about how everyone is “biased” against “conservatives” (though I’d argue the MAGA movement is hardly “conservative”) is really about making sure that anyone in a position of gatekeeping or arbiting gives them more leeway to break the rules, simply to avoid the appearance of bias.
That means that in continually accusing everyone (mainstream media, social media, etc.) of unfair bias against the MAGA movement, we actually get the exact opposite: an unfair bias that gives MAGA folks a pass on breaking not just the rules, but general societal norms like… not contesting the results of a presidential election.
Two years ago (just as Elon Musk was gearing up to acquire Twitter to fight back against what he insisted was “bias” in their moderation policies), we wrote about a preprint of a study by a group of researchers, including David Rand, Mohsen Mosleh, Qi Yang, Tauhid Zaman, and Gordon Pennycook.
This week, an updated version of that study has finally been published in the prestigious journal, Nature. Its findings are pretty clear: content moderation does not appear to be focused on ideology, but does target potentially dangerous disinformation. The simple reality is that the MAGA world is way, way, way, way more likely to post absolute fucking nonsense.
We first analysed 9,000 politically active Twitter users during the US 2020 presidential election. Although users estimated to be pro-Trump/conservative were indeed substantially more likely to be suspended than those estimated to be pro-Biden/liberal, users who were pro-Trump/conservative also shared far more links to various sets of low-quality news sites—even when news quality was determined by politically balanced groups of laypeople, or groups of only Republican laypeople—and had higher estimated likelihoods of being bots. We find similar associations between stated or inferred conservatism and low-quality news sharing (on the basis of both expert and politically balanced layperson ratings) in 7 other datasets of sharing from Twitter, Facebook and survey experiments, spanning 2016 to 2023 and including data from 16 different countries. Thus, even under politically neutral anti-misinformation policies, political asymmetries in enforcement should be expected. Political imbalance in enforcement need not imply bias on the part of social media companies implementing anti-misinformation policies.
I think it’s important that these researchers point out that they even had groups of “only Republicans” rate the quality of the news sources that the MAGA world was pushing.
Often in discussions around bias in a different context, there are debates about whether or not it makes sense for there to be equality in opportunity vs. equality in outcomes. This is often demonstrated in some variation of this graphic, created by the Interaction Institute for Social Change, which has become quite a meme and comes up in lots of culture war discussions.
But, in many ways, the debate on social media moderation and bias is just a different form of that same argument (though, in some weird ways, with the viewpoints reversed from conservative/liberal thinking). On issues of bias in opportunity, the “traditional” (grossly generalized!) view is that “conservatives” want equality in opportunity (the left side of the picture) and “liberals” prefer equality of outcomes (the second picture).
When it comes to social media moderation, the roles seem somewhat reversed. The MAGA world insists that since they get moderated more often, showing that the “outcomes are uneven,” it proves an unfair bias.
Yet, as this study shows, if the inputs (i.e., the likelihood of sharing absolutely dangerous bullshit nonsense) are uneven, then of course the outputs will be uneven.
And that’s even true after working the refs. When the MAGA world is so committed to pushing blatantly false misinformation, some of which could cause real harm which a platform might not want to support, the outcome may still show that they end up getting suspended more often, even when sites like Facebook bend over backwards to give MAGA folks more leeway to violate its rules.
The study makes that clear. It notes that the greatest predictor of getting suspended was not “are you conservative?” but “are you sharing bullshit?” For people who supported Trump but didn’t share nonsense, they were less likely to be suspended. People who supported Biden (in 2020) but did share nonsense, were more likely to be suspended.
The determining factor here was sharing nonsense, not political ideology. It’s just that Trump supporters shared way more nonsense.
The researchers also explore what would happen if a totally “neutral anti-misinformation policy” were implemented. And… they found nearly identical results:
Using this approach, we find that suspending users for sharing links to news sites deemed to be untrustworthy by politically balanced groups of laypeopleleads to higher rates of suspension for Republicans than Democrats… For example, if users have a 1% chance of getting suspended each time they share a low-quality link, 2.41 times more users who shared Trump hashtags would be suspended compared with users who shared Biden hashtags (d = 0.63; t-test, t(8,998) = 30.1, P < 0.0001). Findings are equivalent when basing suspension on expert assessments of the 60 news sites, or when correlating predicted suspension rate with ideology (0.31 < r < 0.39, depending on ideology measure; P < 0.0001 for all); …
[….]
These analyses show that even in the absence of any (intentional) disparate treatment on the part of technology companies, partisan asymmetries in sanctioned behaviours will lead to (unintentional) disparate impact whereby conservatives are suspended at greater rates. From a legal perspective, political orientation is not a protected class in the USA and thus neither form of disparate treatment is illegal (although potentially still normatively undesirable). Although disparate impact may reasonably be considered to constitute discrimination in some cases (for example, employment discrimination on the basis of job-irrelevant factors that correlate with race), in the present context reducing the spread of misinformation and the prevalence of bots are legitimate and necessary goals for social media platforms. This makes a normative case for disparate impact on the basis of political orientation.
This shouldn’t be surprising to folks who have followed this space for a while. Indeed, it confirms a lot of what many of us have been saying for years. But it’s certainly nice to have the data to support the findings.
The Detroit PD is a case study in misuse of powerful surveillance tech. The department is notable for being involved in no less than three wrongful arrests, due to misuse/abuse of its facial recognition tech. The city has already paid out a $300,000 settlement in one of these cases. Worse, two of the three cases involve the same so-called “detective,” which means one of their investigators should definitely never be allowed to use the tech again.
These arrests are the result of ignoring the guidance that surrounds the tech. The tech provider warns facial recognition matches should not be used as the sole probable cause for arrest warrants. The PD’s policies say the same thing. And yet, there have been multiple wrongful arrests because investigators ignored these policies and procedures, including one who’s done this at least twice.
Now, there’s another misuse of surveillance tech that’s making the wrong kind of headlines for the Detroit Police Department. Automatic license plate readers capture hundreds of millions of plate/location images across the country every day. These are automatically searched against hot lists of vehicles tied to criminal activity.
This is an efficiency gain — one that allows officers to more with less legwork. But it also accelerates the error rate, something that has also been linked to wrongful stops and arrests. That the Detroit PD would also be misusing this tech isn’t a surprise. The bigger surprise might be that the Detroit PD specifically allows officers to engage in this misuse. This is from Paul Egan’s article for the Detroit Free Press, which details yet another tech-enabled screwup by the Detroit PD.
A Detroit police spokesperson said officers are authorized to use license plate readers in reverse, when they have a vehicle description but no partial license plate number.
That’s not responsible use of this tech. This is nothing more than a policy-enabled fishing expedition, which allows investigators to surf the database for suspects, using nothing more than speedy guesswork to find a “suspect.”
This case — which has resulted in a lawsuit — is even worse. Investigators didn’t even have a partial plate number. All they had was a vehicle description. So, they googled the haystack until they found a car to stop, assuming whoever was in it must be the criminal they were looking for. They were wrong, but that didn’t stop them from sending a bunch of officers to this person’s house, throwing her 2-year-old child into a cruiser while they cuffed her and, basically, stealing her car.
Acting on information drawn from its multi-million-dollar network of license plate readers, Detroit police handcuffed Isoke Robinson, put her 2-year-old son, who has autism, in the back of a police cruiser, and seized and impounded her only car for more than three weeks.
But witnesses to the Sept. 3, 2023, drive-by shooting police were investigating never gave officers even a partial license plate number.
And the man later convicted in the nonfatal shooting has no known connection to Robinson or her 2013 Dodge Charger. Robinson, who had to borrow someone’s truck to keep her job as an assembly line worker at Stellantis and has sued the city in federal court, was never charged with a crime or even identified as a suspect.
Based on nothing more than a reverse image search (so to speak), officers took Robinson off the road, out of her car, and traumatized her son. The testimony given by officers in the lawsuit makes it clear no one was handling anything professionally or responsibly. The plate reader cops looked at was two miles from the crime scene but only a couple of blocks from where Robinson lived. (She has since moved out of Detroit.)
On top of that, images of the suspect vehicle obtained from nearby private CCTV cameras showed something an actual investigator might have attempted to verify before prematurely declaring victory over crime just because they’d take an innocent woman’s car. Those images showed a Dodge Charger with only a single working fog light. Despite holding onto Robinson’s car for three weeks, no one bothered to take another look at the car cops were carelessly claiming was evidence.
Detroit Police Detective Dion Corbin, Jr., who was in charge of the shooting investigation, testified in an Aug. 14 deposition that the entire time police had Robinson’s vehicle impounded, they never checked to see whether both her fog lights were working, or not. They also never tested the car for potentially relevant evidence, such as gunpowder residue, Corbin testified.
These people don’t deserve the title of “detective” or “investigator.” They’re doing neither of these things. They’re just finding the simplest, most effort-free “solution” to their problem. And, by doing so, they’re neither finding criminals, nor are they making Detroit any safer. While they’re impounding cars and throwing toddlers into cop cars, the real criminals are still out there, avoiding arrest by doing nothing more than existing alongside officers who can’t be bothered to take their own work seriously.
The Complete ChatGPT Artificial Intelligence OpenAI Training Bundle has 4 beginner-friendly courses to help you become more comfortable with the capabilities of OpenAI and ChatGPT. You’ll learn how to write effective prompts to get the best results, how to create blog posts and sales copy, and how to create your own chatbots. It’s on sale for $30.
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.
In recent weeks there has been a flurry of laws, regulatory proposals, and lawsuits regarding “deepfakes,” along with the usual rising levels of concern in the media about how the world won’t be able to handle this. For some perspective, the Pessimist’s Archive just published a story highlighting how a nearly identical fear gripped the world in 1912 regarding “fake” photos. We’ve republished it here, with permission, though you really should subscribe to the Pessimist’s Archive as well.
Concern about deceptively edited photos feels like a very modern anxiety, yet a century ago similar worries were being litigated…
Portrait photography gave rise to an industry of photo ‘retouching’ – analog ‘beauty filters’ – to flatter subjects in a way portrait painters once did. This trend lead to questions about technology distorting our perceptions of beauty, reality and truth:
An 1897 issue of the New-York Tribune would declare the assumption “Photographs Do Not Lie” an “exploded notion”, saying:
Other commercial applications of photo retouching emerged: in 1911 tourists visiting Washington D.C. could acquire fake photographs of themselves posing with then President of the United States William Taft. This troubled government officials. Upon discovering the practice in 1911, a United States Attorney ordered it stopped:
One (literal) photo-shop offering the novel souvenirs appealed to the White House for its blessing to continue the trade, but was denied. (the practice was not against the law, but pressure from the White House appeared effective if only temporarily — in the capital)
Photographic Crimes
The following year a fugitive – wanted for people trafficking – was found in possession of a fake photo posing with President Taft, it was reported he’d used it to buy the trust of his victims:
That this seemingly benign practice had been weaponized prompted some to demand it be regulated against abuse. The Justice Department prepared a law, that was introduced by then Senator Henry Cabot Lodge — who’d similarly been troubled after reportedly finding a photograph of himself with someone he’d never met.
On July 29, 1912 the bill was introduced to the Senate: “to prohibit the making, showing or distributing of fraudulent photographs”:
The law would make it illegal “to make, sell, publish, or show” any “fraudulent or untrue photograph, or picture purporting to be a photograph” of anyone who had not first given permission. Violation would see punishment of up to 6-months in jail or up to a $1000 fine ($31,797 adjusted for inflation.)
The proposed rules made headlines across the US, bringing the topic of fake photographs to wider attention:
Not everyone was a fan of the proposed law however. Some representing the photography community worried about unintended consequences of the bill: a 1912 edition of the publication ‘American Photography’ called the bill ‘indefensible’, saying that the law would make photographers and publishers “continually liable to blackmailing suits” – we assume by those claiming to have never given permission.
It appears the bill was never voted on and died in the Senate. The following year – President Taft would lose an election to Woodrow Wilson – reports of a renewed trade in fake photos of President Wilson in the capital would soon appear.
The Times Leader Wilkes-Barre, Pennsylvania • Fri, Mar 7, 1913 Page 3
What if…
Had the law passed it would have been very relevant in the 21st Century, with the rise of digital photography, photoshop and more recently AI enabled image editing.
It would have made a federal crime of removing an object from a photo for instagram – without permission of everyone in it. On the other hand, recent efforts to legislate against fake nude images would be unneeded – those would be illegal already.
Surprisingly, this latter issue was also present in prior centuries: unmentioned in defense of the proposed law, were real cases of ‘photographic slander’ against women: in 1905 a gang was using the threat of “the circulation of indecent trick photographs for the purposes of blackmail.” A 1891 report out of Chicago noted the arrest of a “gang of scamps” selling fake nude images of high society women. (1936 would see a blackmail plot against opera stars)
Fake nudes could have been banned over a century ago – but the law was too broad, focusing on more speculative, rather than specific proven harms. A similar critique is made of some AI regulation proposed today – making this, a good possible lesson for the best way to approach AI regulation.
Side Note: Hypocrisy
A year after the law was proposed – in 1913 – a photo would appear of President Taft atop a Carabao – the national animal of the Philippines. It was thought to have been part of an effort to buy goodwill with a nation seeking independence from the United States:
It turns out, it was the very kind of fake photo the Taft administration was railing against. In 2018 history researcher Bob Couttiediscovered a 1908 photo of President Taft astride a horse, in an identical pose: comparing the images it seems undeniable Taft had been cut and pasted onto a different image: literally.
Secretary of War William Howard Taft on a horse, 1908. Via New York Times Archives
We’ve noted repeatedly how in 2020, the Trump administration tried to give Elon Musk’s low-Earth orbit (LEO) satellite broadband company Starlink nearly a billion dollars in taxpayer subsidies to connect some traffic medians and airport parking lots.
We’ve also noted that Musk and his Republican friends have been whining like toddlers because the Biden FCC backed away from those subsidies, correctly noting that they weren’t sure that the expensive and increasingly congestion-plagued satellite service could consistently deliver quality speeds. Instead, the Biden FCC (again, correctly) argued that if you’re going to spend billions of taxpayer money on broadband, it should be on more reliable and less congested options like fiber, 5G, or fixed wireless.
Enter Hurricane Helene, which recently carved a swath of destruction from Florida into the Carolinas, destroying homes and knocking out power and cell service for millions of people. Musk apparently couldn’t help but inject himself into the story, using the suffering of Helene victims as the perfect opportunity to whine that he should have gotten more taxpayer money:
If you can’t see it, that’s Musk claiming over at his post-truth network that if the FCC had agreed to give Starlink a billion dollars, there’d be nearly 20,000 working satellite kits in areas impacted by Helene. He goes on to falsely accuse the FCC of killing people and breaking the law, adding “lawfare costs lives.”
There was nothing illegal about the FCC’s decision. Even if Musk’s Starlink had received the money, the build out the money was funding wouldn’t have started until 2025. And even if Starlink had received that money, there’s no guarantee that locals could have afforded the expensive $120 a month (plus hardware fees) the service costs. Or that the power would have been on so that the dishes could have been used. Or that roof-mounted dishes would have survived the storm.
“Chairwoman Rosenworcel stands by the FCC’s thorough review of a program meant to provide long-term access to reliable and affordable broadband in rural communities,” the commission said in a statement provided to PC Magazine. “In this instance, the agency denied public funds to more than a dozen companies—not just Starlink—who did not meet the program requirements. As an independent agency, the FCC takes seriously its obligation to ensure that taxpayer dollars only go to entities that fully comply with the rules and the law.”
When it comes to subsidizing broadband, you want taxpayer money going to the most beneficial and cost effective option. Usually that’s something like middle mile or open access last mile fiber networks, which not only drive high-capacity, “future proof” fiber connectivity into a region, but also improve local cellular connectivity, and lower the cost of market entry to boost regional broadband competition.
If you can’t do that, you subsidize local 5G and fixed wireless efforts, which are more congested than fiber but still generally more reliable and less capacity constrained than satellite. Starlink is a fine option to fill in the coverage gaps after you’ve done all of that, but again it lacks the capacity to truly scale and it’s expensive — which is a problem given that cost is the biggest obstacle to U.S. broadband access at the moment.
What you don’t do is throw billions of unaccountable taxpayer dollars at a petty billionaire’s expensive satellite venture that may or may not even exist ten years from now, and is increasingly seeing congestion and slowdowns due to over-saturation of the physics-constrained network.
Of course Musk being Musk, he couldn’t help but inject himself into a massive tragedy to try and generate pity and grab additional taxpayer money. Meanwhile, the FCC is actually coordinating disaster response and tracking service and 911 outages, while providing discounted communications access to impacted survivors. FEMA says it’s also trying to leverage all manner of options, including Starlink, to shore up regional emergency connectivity.
To be clear Musk did do more than tweet. The conspiratorial CEO claims he did actually send some Starlink terminals to the powerless region, though Musk being Musk, and the U.S. press being the U.S. press, it’s not entirely clear if anybody actually confirmed delivery. Most of the press ignored his gross subsidy quip, and immediately jumped to portraying Musk as an altruistic life saver who is single-handedly restoring connectivity to the region:
Usually what winds up being the most help in disasters like this (after food and water) are good old traditional HAM radio operators, who help orchestrate basic coordination when all other options are out of commission. What doesn’t necessarily help as much are loud-mouthed, petulant and conspiratorial billionaires with more money than sense, keen on exploiting tragedies to undermine government regulators and glom on to taxpayer subsidies they routinely pretend to have disdain for.
