Escort Gigi Lenoir recently went viral for a chilling warning on social media. A friend of hers in the sex work industry had been pulled aside at the U.S. border, shown printouts of her website, and banned from the U.S. for five years. “Ladies that are face out, be careful. Facial recognition is out of control,” Lenoir wrote, referring to sex workers who include pictures of their faces in public advertisements. The tweet went viral and provoked a conversation among sex workers about digital safety.

Omg one of my friends was detained at the border and DENIED entry to the USA for 5 years. She used to be face out and they basically had her whole website printed out and interrogated her. Ladies that are face out, be careful. Facial recognition is out of control

— Gigi Lenoir BIRTHDAY MONTH (@exploregigi) July 13, 2026

As a former State Department official who handled immigration, I have been trying to find out if the government is really using facial recognition to find the identity of foreigners who have done sex work online. An American sex worker told me a story of her border interrogation that suggested she was flagged by a computer program. When I asked, a spokesperson for Customs and Border Protection (CBP) didn’t deny that the agency was using facial recognition to identify sex workers.

Stories of new forms of border harassment under the second Trump administration are a dime a dozen these days. Some of them are frightening or depressing because of their absurdity or sheer cruelty. Others seem to elide a bigger structural change. But the administration’s approach to sex work is in line with longstanding and confusing U.S. policy.

U.S. immigration law’s perspective on sex work is not about exploitation, harm, or even any coherent standard about what is legal and what isn’t. While the law bans people from coming to the country under a vague definition of immorality, it also allows authorities to give special talent visas to popular OnlyFans models. Ordered to crank up enforcement on every metric, and given new tools of dubious utility, officers start going after easy targets: unpopular radicals and “low” women.

The American sex worker I spoke to was a dual citizen returning home to the U.S. from the other country where she holds citizenship. As a U.S. citizen, there was no way the government could deport her. But she was pulled into secondary screening at the border and could not get a straight answer for why she was being held. The other people in the waiting room were single Middle Eastern men, families and couples from Latin America and Africa, and women with a “bimboesque, influencer look” from her country, she says.

After a long wait, a man in a suit questioned her without identifying what agency he was from. Despite her U.S. citizenship, the man insisted on speaking in her other country’s language, as if he was testing her fluency. He searched her bags and asked about her travel patterns, source of income, and exact addresses. Based on my experience in government, what happened at the border suggests that she had been flagged as a sex worker by a computer program.

CBP doesn’t deny that it’s doing this, though it didn’t definitively confirm it either.

“U.S. Customs and Border Protection uses technology—including biometric tools—to support lawful border security operations, verify traveler identity, and facilitate legitimate travel consistent with applicable law and policy. Any inspection or enforcement action at the border is based on the totality of the circumstances and conducted consistent with CBP authorities, privacy requirements, and civil rights and civil liberties protections,” a CBP spokesperson told me. “Under the Immigration and Nationality Act, an alien who is coming to the United States to engage in prostitution, or who has engaged in prostitution within 10 years of the date of application for admission, may appear to be inadmissible to the United States.”

This isn’t a situation in which the U.S. government is looking at combating sex trafficking or sexual exploitation seriously or systemically. If anything, the Trump administration is pulling officers off sex trafficking cases to do other immigration enforcement work.

There are several overlapping and vague definitions of “prostitution” and “commercialized vice” in U.S. immigration law, based on a perceived threat to public morality or the idea that sex workers themselves are bad people who shouldn’t come here. They apply even to people who did fully legal sex work in their own countries. The same immigration system is fine with the production of pornographic media, treating workers in those industries like any other entertainer, unless their work crosses the line into “commercialized vice,” a term that is undefined.

This vagueness has led courts to make sex worker entry bans easy to overturn. One of the most famous cases in immigration law, Matter of Hranka, was about a Canadian citizen deported from the U.S. in 1975 for engaging in sex work in Detroit. The Board of Immigration Appeals ruled that the government should grant visitor waivers for people with this kind of past on even the most minimal grounds.

Ironically, the system has actually started rewarding the most successful OnlyFans performers (among other social media influencers) with O-1 visas for “extraordinary ability.” Some immigration lawyers told the Financial Times earlier this year that social media influencers make up more than half their clientele, simply because followers and likes are easy metrics that can be used to prove “talent” as an entertainer. The woman who described her border experience to me took it as par for the course for the realities of sex work and the hypocrisy of government enforcement.

“Girls that are high-end and citizens serving exclusively rich clientele have a way different experience to girls working the streets to live,” she says.

The woman also described the conversations that sex workers have as they struggle to figure out the limits of digital safety.

“A lot of these girls are really just mostly normal girls following the formula set before them by other girls. They blur their face out because it’s what all the other girls do, and to some degree, it might offer some protection for sure, say, from a stalker maybe,” she says. “The government and its tools are a different game. The tools measure faces’ proportions, even with blur. And most girls have nowhere near a strong enough blur to offer any real protection.

“A lot of girls have OnlyFans, Fansly, ManyVids,” she adds. “That’s literally tied to an ID. I think as a sex worker, to have any real sense of privacy, you have to be face out entirely, not even blur, no tattoos, and only operate via secure channels/bitcoin and no selling on normal platforms.”

These vulnerabilities are built into having any online presence at all, and it’s difficult to stay anonymous in a world where new global laws are attempting to tie online presence to verified ID, ostensibly to prevent sexual exploitation. After all, a hypothetical sex worker in Europe who wants to start an OnlyFans account might have de-anonymized themselves to the U.S. government in the process of doing so. It’s unsettling to think of all the people being tracked by the government of a country where they do not live and do not plan to work.

The U.S. immigration system functions in an increasingly inchoate way. It isn’t clear what’s legal or not, or what kind of information the government is gathering on people. The chaos means different standards for different people, benefiting the most wealthy and well connected, as the O-1 visas for OnlyFans models demonstrate. That’s the greatest irony of all. Instead of making border control fair and forthright for everyone, the rise of automatic systems and algorithms is making it more arbitrary and biased.

The post Customs and Border Protection Doesn't Deny Using Facial Recognition To Target Foreign OnlyFans Models appeared first on Reason Magazine.

The post Today in Supreme Court History: September 19, 1907 appeared first on Reason Magazine.

States’ efforts to regulate election deepfakes keep colliding with the First Amendment.

On Wednesday, U.S. District Judge Susan Watters issued a preliminary injunction preventing Montana from enforcing a 2025 law that restricts AI-generated political advertisements. The law had been challenged by Accountability in State Government, a group run the former state lawmaker Dan Bartel; the organization had sent mailouts containing doctored images of political figures holding pride flags and wearing pronoun badges. 

Under Montana law, it is illegal to sponsor or distribute an election communication within 60 days of voting if you “knew or should have known [it] is a deepfake”—though “satire or parody” is exempt. And so the targets of Bartel’s mailouts filed complaints. State Rep. Jennifer Carlson (R–Manhattan), for example, said a mailer showing her holding an AI-generated pride flag was created “with the intent to injure the reputation of the candidate or otherwise deceive a voter.”

The complaints were dismissed on August 20, with Christopher Gallus, commissioner of political practices, concluding that the mailers were sufficiently “humorous or outlandish” to qualify as satire. The next morning, Watters heard arguments on Bartel’s request to block enforcement of the law altogether on the grounds that it violates the First Amendment.

This week, Watters found that Bartel and his group are likely to succeed on their First Amendment challenge. Montana’s statute covers synthetic media created “with the intent to injure the reputation” of a candidate or political party, meaning that, as Watters put it, the law “treats identically deceptive media differently based solely on whether its message is weaponized to damage or deployed to elevate a candidate.”

Adam Schulman—an attorney at the Hamilton Lincoln Law Institute, which is representing Bartel in the case—says that this was the statute’s fundamental flaw. “You could have AI deepfakes that were fine as long as they were complimentary,” Schulman tells Reason. “You could have Donald Trump dressed in a Pittsburgh Steelers uniform deepfake [and that would be] fine because it’s not injuring the reputation of a candidate.”

Schulman also argues that the law asks regulators to make inherently subjective judgments about political imagery. Montana’s statute relies in part on how a “reasonable person” would understand the material, but that becomes difficult when satire and parody are involved. “People have different artistic sensibilities,” says Schulman.

Montana is not alone. Plenty of states have attempted to restrict AI-generated political speech and discovered the constitutional difficulty of doing so. After the conservative YouTuber Christopher Kohls posted an AI-produced parody of then–Vice President Kamala Harris in the 2024 election, California passed Assembly Bill 2839, prohibiting the distribution of communications “that contains certain materially deceptive content” produced “with malice.” In 2025, a federal court ruled the law was unconstitutional and permanently enjoined California from enforcing the law.

Similarly, Hawaii passed Act 191 in 2024, which restricted “materially deceptive” election media. The Babylon Bee and Dawn O’Brien, a Hawaiian resident, challenged the law as a violation of free speech. In January 2026, U.S. District Judge Shanlyn Park granted a summary judgment finding the law “unconstitutionally vague,” in part because it “employs ambiguous and subjective standards.”

As new technologies develop, calls for more government regulation inevitably ensue. The push to regulate AI-generated political speech reminds Schulman of COVID-era efforts to police “misinformation.”

Montana has a month to appeal the decision. But given how things have gone for other states, it might not want to bother. The constitutional limits on regulating AI-generated political speech may not be fully settled, but the early cases have been a warning against attempts to restrict protected political expression.

The post Montana's Anti-Deepfakes Law Just Hit a First Amendment Roadblock appeared first on Reason Magazine.

Is that constitutional? According to the D.C. Circuit (the federal appellate court with jurisdiction over D.C.), it’s

• unconstitutional for the government to exclude press outlets based on viewpoint from spaces that have been broadly opened to the press, such as the main White House press briefing room but
• constitutional to do so as to spaces that aren’t generally open to the press, such as interviews in the Oval Office.

The key precedent on this in AP v. Budowich (D.C. Cir. 2025), which upheld the exclusion of the AP from the Oval Office (because of AP’s refusal to call the Gulf of Mexico “Gulf of America”), but reaffirmed (citing Sherrill v. Knight (D.C. Cir. 1977)) that

When the White House opens its facilities to the press generally, as it does in the Brady Briefing Room, it cannot exclude journalists based on viewpoint. Sherill v. Knight.

To be sure, if the exclusion is based on some objective viewpoint-neutral judgment about which outlets are so factually inaccurate that they don’t merit inclusion, or about the $8.2M in 2024 federal government subscriptions to Politico newsletters, that might be permissible. But if the exclusion is based in part on disapproval of CNN’s, MS NOW’s, and Politico’s perspectives or editorial judgments as being unfair or unsound, that would likely be seen as viewpoint-discrimination—which, again, is unconstitutional as to the main press briefing room but constitutional as to the Oval Office.

The post Can President Trump Ban CNN, MS NOW, and Politico from White House? appeared first on Reason Magazine.

A failed Senate procedural vote on the Digital Asset Market Clarity Act on Tuesday may have saved crypto consumers from a questionable regulatory bill backed by industry insiders. 

The bill—a first-of-its-kind regulatory framework for the crypto industry—would have drawn clear jurisdictional lines between the Securities and Exchange Commission (SEC) and the Commodity Futures Trading Commission (CFTC) on regulating the offer and sale of digital commodities.

Instead of the SEC regulating crypto through individual enforcement—which it currently does—the bill would give Americans holding digital assets a pathway to legal certainty by establishing disclosure requirements, preserving both agencies’ anti-fraud authorities and creating risk-management, cybersecurity, and compliance standards for decentralized finance. 

Still, the bill also gives federal agencies far too much discretion, authorizing them to set new regulations and exemptions and coordinate with international regulators if it is “in the public interest or for the protection of investors,” an invitation for career bureaucrats to decide what’s best for Americans.

After a year of bipartisan negotiations, The Wall Street Journal reports the bill failed to pass a procedural vote—which would have allowed it to be fully voted on by the Senate—because Democrats weren’t sold on the bill’s ethics provisions as strong enough to prevent President Donald Trump “from continuing to profit off digital assets.” 

It’s a statement that seems to ring true, considering an ABC News report that Trump’s June financial disclosure reports showed crypto earnings of more than $1.4 billion. This includes more than $591 million from World Liberty Financial (WLF), $636 million from his meme coin business CIC Digital, and more than $196 million from the equity sale of a holding company linked to WLF. Last year, the president announced that he would hold an “intimate private dinner” for the top 220 holders of his $TRUMP meme coin. In the days after the announcement, the coin’s value surged by 50 percent.

The bill would bar the president, vice president, members of Congress, senior federal officials, and their spouses from profiting from crypto tokens, promoting tokens, or owning a meaningful stake ($15,000 or more) in a crypto venture such as WLF. Anyone who currently owns a stake would have to sell it or place it in a blind trust—or face fines of at least $500,000 and the potential loss of any profits.

These provisions are a step in the right direction, but they’re largely milquetoast and do little to address the air of corruption surrounding the president’s crypto dealings. 

With a blind trust, the president could still profit from WLF, and any new tokens the company launches would be grandfathered in under the bill. The ethics provisions also fail to account for children and other relatives. That would seemingly insulate WLF, which was launched by Trump’s sons and the sons of U.S. peace envoy Steve Witkoff. The bill tasks the Department of Justice (DOJ) with enforcing penalties against violators. Given what we’ve seen from the department under Trump, it’s fair to wonder whether the attorney general would remain impartial.

The bill’s enforcement mechanism gives state attorneys general the power to sue the Justice Department for any harm to a state or its residents. However, it does not allow them to sue federal officials. State AGs are also empowered to hold trading platforms that run afoul of state laws accountable.

Despite the disagreements, the bill is still on the Senate calendar and could resurface. On Tuesday, Sen. Thom Tillis (R–N.C.) said this is “not the end” for the bill. It might be wishful thinking from Tillis; Forbes reports that it’s “unlikely Senate Majority Leader John Thune will allocate calendar time without certainty that they have enough Republican and Democratic commitments.”

Lawmakers’ back-and-forth hasn’t stopped the push to regulate the crypto industry. 

On Thursday, in the absence of congressional leadership, the SEC issued an order—part of the agency’s Project Crypto initiative to make America the center of the crypto industry—allowing certain trading platforms to issue digital tokens that represent publicly traded U.S. stocks. The order grants a five-year Innovation Exemption to platforms that facilitate tokenized stock trading and further integrates digital assets with traditional markets like Nasdaq and the New York Stock Exchange. 

In a statement on X, CFTC Chairman Mike Selig called the failed vote on the bill “unfortunate,” adding that the CFTC is “locked in and ready to ship its rules for the new frontier of finance.”

The post Trump's $1.4 Billion Crypto Haul Sinks the Senate's Crypto Bill appeared first on Reason Magazine.

President Donald Trump’s habit of ostracizing America’s allies is making it a precarious time to be American, European, or Canadian.

European Commission President Ursula von der Leyen summarized this precarity at Wednesday’s State of the European Union address, saying, “We must urgently reimagine our partnerships.” Though the address was delivered to the European Parliament, the remarks were directed to Canadian Prime Minister Mark Carney, who was visiting the assembly and conspicuously seated to von der Leyen’s left on the stage.

Von der Leyen then invited Canada to join the European Union (E.U.) as the bloc’s first “associate member.” The designation does not yet have a clear definition, but the two sides have said that they hope it will yield closer cooperation in energy, trade, technology, and Arctic security. In his speech to the European Parliament on Thursday, Carney said he intends to integrate the two markets further and join the E.U.’s student exchange program, Erasmus+.

Canada and the E.U. have long been partners, and the trend is accelerating. Under the 2017 Comprehensive Economic and Trade Agreement (CETA), tariff barriers on most trade between the two economies dropped. Earlier this year, Canada joined the SAFE Instrument, an E.U. program for joint defense rearmament. The country is also a cooperative participant in the European Space Agency, the E.U.’s research fund Horizon Europe, and E.U. election monitoring missions.

Details of the arrangement and how Canada will integrate into the E.U. have yet to be released. It might resemble the bloc’s existing agreement with Norway, Iceland, and Liechtenstein called the European Economic Area, which integrates the three states into the E.U. common market, and allows for the free movement of goods, capital, and labor between them. Bilateral agreements with Switzerland have similarly removed trade and travel barriers, reports Bloomberg. In both cases, these “partial members” have acceded to E.U. regulations without joining E.U. institutions, in exchange for retaining autonomy to regulate the fishing and agriculture industries, and write their own trade policy. Though much space for further integration between Canada and the E.U. is possible, experts say that Canada joining the European Union Customs Union or adopting E.U. laws is unlikely.

Details aside, the political message is clear: Canada and the E.U. are seeking closer relations as they are driven away from their traditional ally, the United States. And honestly, who could blame them? 

Since retaking office, Trump has wielded aggressive rhetoric and tariff policy against the E.U. and Canada. 

On April 2, 2025, Trump announced his “Liberation Day” tariffs, which set a baseline 10 percent tariff rate on imports from most countries, with a 20 percent rate applied to E.U. member states. In July 2025, after months of threats and negotiations, Europe and the U.S. reached a loose trade agreement that subjected most imports from the E.U. to a 15 percent tariff, while tariffs on U.S. industrial exports were dropped. The E.U. also agreed to “invest some $600 billion in the United States and dramatically increase its purchases of U.S. energy and military equipment,” reports Reuters. This agreement did not stave off the president, and over the following year, Europe saw further threats as the deal was being fully ratified and implemented. Von der Leyen’s speech seems to have particularly invigorated Trump, who is now threatening to impose “serious tariffs” on the E.U. if Canada joins the bloc.

Though Canada was spared from Liberation Day’s tariffs, sector-specific import taxes have steadily expanded over the past two years. After a rocky summer of U.S. reticence and threats, trade negotiations over Trump’s proposed 50 percent tariff on several Canadian imports collapsed, and the tariffs went into effect last month. The conflict has since escalated into a full-scale trade war, with Canada applying retaliatory tariffs and consumers in both nations suffering the cost. Trump’s jingoistic attitude toward a neighbor—frequently calling for Canada to become the “51st state,” and recently ordering the renaming of Lake Ontario to Lake America—has not helped to de-escalate the situation.

The administration’s tariff policy against two of America’s largest trading partners has never made sense, nor has it shown much of a coherent strategy. The initial legal justification for the Liberation Day tariffs was a decades-old law that grants the president emergency powers, which plainly has nothing to do with tariffs. The Supreme Court struck down this justification in February, but swift lawyering ensured that new mechanisms for enforcement were found. The prevailing uncertainty over trade policy, coupled with the administration’s contradictory rationale for these duties (Were they imposed to bring back manufacturing? Or reduce the trade deficit?) has resulted in whiplash for negotiators and businesses. It’s also led U.S. allies to realize—not unfairly—that Trump’s trade agenda has always been more about aggrandizement and power than sound economic policy.

Carney said as much during his speech to the E.U. Parliament on Thursday. “[Europe and Canada] do not seek power to dominate others. On the contrary, we are pursuing resilience so that no one can control our open markets, impair our sovereignty, threaten our territorial integrity, or undermine our freedoms, our democracies, our rule of law.”

The post Canada Is Moving Closer to Europe. Trump's Tariffs Help Explain Why. appeared first on Reason Magazine.

The “safety” of artificial intelligence is the talk of the town in Washington. Over the summer, the companies OpenAI, Anthropic, and Meta all disclosed that their software had hacked into outside computers of their own volition. Most worryingly, OpenAI’s virtual “agents” coordinated as a swarm to cover their tracks. Last week, Anthropic researcher Jacob Coxon publicly resigned out of his worry that computers will become “smart enough to kill us.” While President Donald Trump dismissed the fears, Sen. Bernie Sanders (I–Vt.) proposed jailing AI developers.

Across the ocean, another AI safety controversy was unfolding. Israeli journalists Yuval Abraham and Rachel Szor premiered their documentary NAZA, named after the Hebrew acronym for collateral damage, at the Venice Film Festival in Italy last week. Building on Abraham’s earlier reporting on his country’s targeting practices, the film airs testimony from Israeli intelligence officers about how they used AI to surveil and kill Palestinians in Gaza, including civilians. The Israeli military, which denies the allegations, has launched an investigation into the “leaking of classified materials” to Abraham and Szor, according to The Times of Israel.

Israel isn’t the only country using AI to mark people for death. The Russian and Ukrainian armies have both been rolling out autonomous drones, which use AI guidance software to choose targets when they are cut off from radio contact with pilots, over the past few months. The U.S. military itself used an AI system to select over 1,000 targets in the opening surprise attack on Iran. On the other side, the Iranian military and its allies in Yemen used Anthropic’s Claude to gather intelligence and develop weapons guidance software, according to a recent security report by Anthropic.

“The future is already here—it’s just not evenly distributed,” cyberpunk science fiction writer William Gibson famously said. Many people around the world now live in AI kill zones, where inhuman algorithms summon robot assassins against anyone in the wrong place at the wrong time. And unlike Coxon and Sanders’ warnings, the danger isn’t in the nature of AI itself. Instead, this new technology is a tool empowering people to do what they long wanted to do. In the case of governments, that means spying on and bombing their enemies with fewer and fewer constraints.

Take the massacre of schoolchildren in Minab, Iran. During the opening attack on Iran, the U.S. military bombed Minab’s elementary school, which used to be an Iranian navy building, killing 150 people. Some journalists initially speculated that AI was responsible for identifying the school as a target. Although the military has been tight-lipped about its investigation, leaks suggest that human negligence had a major role: Officials ignored warnings from the computer that their intelligence was old, and an analyst had actually updated the building as a school in a separate database that wasn’t connected to the targeting system.

In other words, AI didn’t make the decision to kill those children. At most, it sorted through data provided by humans and gave a recommendation to humans, who chose to act on it without double-checking their findings.

The value of AI for militaries is being able to integrate large amounts of surveillance data more closely and quickly into what the U.S. Army calls the “kill chain.” Mohammed R. Mhawish, a Palestinian journalist from Gaza, described the experience of life under military rule for New York Magazine: “drones that hovered endlessly overhead, quadcopters that dipped near windows and entered houses, facial-recognition scans at checkpoints, movements followed through phone tracking, calls that broke with static before an air strike.” Human beings can only process a small fraction of that information by hand, and might not be able to pass it on quickly enough to a soldier or drone operator.

The rise of commercial AI has allowed rivals to catch up simply by buying consumer software. Last week, Anthropic released a report on its progress “detecting and countering misuse of AI.” The company caught an “Iran-nexus threat actor that used Claude to collect and analyze publicly accessible data to develop targeting recommendations against U.S. naval forces in the region.” (In other words, Iran was developing a poor man’s Maven.) The report also includes non-state guerrillas, including “a cell of threat actors based in northern Yemen” that asked Claude for feedback on its missile targeting software, and a rebel group against the Iranian government that compiled “psychographic dossiers” on potential recruits.

Computer-assisted targeting began before electronic computers fully came into their own. Nazi Germany infamously used International Business Machines’ punchcard system to identify “enemies of the Reich” and manage concentration camps. During the Vietnam War, the CIA also used International Business Machines’ computers to list and hunt suspected communist infiltrators through the Phoenix Program.

The beginning of modern AI targeting was the SKYNET algorithm, named for the killer AI in the Terminator series. In 2014, former National Security Agency (NSA) Director Gen. Michael Hayden told an audience at Johns Hopkins University that “we kill people based on metadata.” A year later, The Intercept published a document from NSA whistleblower Edward Snowden describing how the agency was training SKYNET to identify Al Qaeda’s “couriers” from cell phone location data.

The U.S. military openly began using AI targeting in 2017. With an initial budget of $70 million and an initial training data set of 150,000 images, the Pentagon announced Project Maven. “People and computers will work symbiotically to increase the ability of weapon systems to detect objects,” Col. Drew Cukor told a conference. The military bombed its first AI-selected target at a testing range in North Carolina in 2020, then used the same software to manage the U.S. evacuation from Afghanistan and provide targets to Ukrainian forces. The surveillance contractor Palantir received a $480 million contract in 2024 to run the present Maven Smart System.

“It’s not Terminator. The machines aren’t making the decisions, they’re not going to arise and take over the world,” Col. Joseph O’Callaghan told Bloomberg in 2024.

But some armies are giving full autonomy to killer machines. Libyan forces fighting in their country’s civil war carried out a drone attack with “fire, forget and find” software in 2020, according to the United Nations, marking the world’s first autonomous attack. Two years later, Russia invaded Ukraine. The battlefield became saturated with unmanned aircraft, and both sides began using heavy radio jamming to cut off drones from their pilots. AI presented a solution: Let the drones pilot themselves.

The Ukrainian drone manufacturer Aero Center claims to have killed Russian troops with an experimental “Terminator mode” for the first time in 2024, which appeared to be a one-off experiment, according to Ars Technica. Former Ukrainian Defense Minister Mykhailo Fedorov told The New York Times earlier this year that Ukraine has run several more recent autonomous drone tests against enemy depots in Crimea. Meanwhile, Russia launched its first known autonomous drone attack in July 2024, killing three civilians at a gas station. Ukrainian officials found the drone’s computer chip intact and discovered that it was likely programmed to target propane tanks, according to The New York Times.

Autonomous weapons are a step closer to the apocalyptic future that many AI opponents fear. But that apocalypse may look less like Terminator, in which an AI being consciously decides to wipe out humanity, and more like the 1983 movie WarGames, in which a military AI program can’t tell the difference between a simulation and reality. The Anthropic hacking incident from earlier this summer had an uncanny resemblance to the plot of WarGames. An AI program, told to perform a hacking test within a simulation, was accidentally hooked up to the real internet.

Mistakes like that can indeed have catastrophic consequences, as the Minab massacre demonstrates. Military AI working as intended can also create a terrifying world, if its users are tasking AI to carry out malicious purposes. Both these scenarios point to a problem of human, not robot, nature. Artificial intelligence is a tool that makes its users more powerful—for better or for worse—and it’s up to other people to rein them in.

The post AI Is Already Killing People—and Governments Are To Blame appeared first on Reason Magazine.

In the 1930s, there were Nazis in America. In the summer of 1934, residents of Griggstown, New Jersey watched with alarm as tents emblazoned with swastikas were erected on the banks of the Delaware and Raritan Canal. Two hundred boys ranging in age from eight to sixteen, dressed in khaki shirts, black rubber-soled shoes, and brown stockings rolled below the knees, conducted military-style drills before swastika flags that were four feet long. These exercises were part of a youth summer camp called Wille und Macht (Will and Might), sponsored by the Friends of the New Germany, the largest Nazi organization in America, with more than ten thousand members.

Camp Wille und Macht opened shortly after the Friends of the New Germany held a massive rally attended by twenty thousand people in New York’s Madison Square Garden. Nazi storm troopers stood at attention atop a red-draped platform. On each side of the stage were black columns decorated with the German eagle bearing the swastika. The Nazis wore swastika armbands and uniforms consisting of military trousers, Sam Browne belts, high boots, and white shirts. Seven hundred policemen patrolled the Garden, attempting to rein in rowdy counterprotesters including the Young Communists’ League and the “Anti-Nazi Minute Men.”

The presence of Nazis in the United States led to widespread calls for hate speech laws. It set off an intense dialogue around the question: Does the First Amendment provide freedom of speech to those who would, if in power, destroy freedom?

During the 1930s, more than eight hundred fascist groups gained a toehold in the United States during the national crisis of the Great Depression. With names like the Silver Shirts, the Citizens Protective League, and the Patriotic Sons of America, the fascist groups, with their promises of hope and prosperity, appealed to thousands of disaffected and impoverished Americans. Some of the groups were native in origin, while some of them, such as the Friends of the New Germany, were Nazi organizations funded by Berlin.

All the groups were relentless in their scapegoating of minorities, especially Jews. In the 1930s, antisemitism was more virulent than at any point in American history. More than three hundred publications in the United States spewed Hitler’s racial theories. Jews were routinely blamed for the nation’s financial woes. Some called the New Deal a “Jewish plot.” Father Charles Coughlin, known as the “Radio Priest,” commanded thirty million listeners at the height of his popularity in the mid-1930s. Coughlin also published a magazine called Social Justice, which was known to print copy verbatim from The Protocols of the Elders of Zion.

Although the fascists never posed a threat to the nation’s security, they had an ominous presence. In New York and New Jersey, Nazis affiliated with the Friends of the New Germany held rallies and marched in the streets chanting “Heil Hitler.” They distributed antisemitic literature and showed Nazi propaganda films from Germany. Everywhere they went, they were met with violence; they were pelted with eggs, and hecklers cut off their speeches. Several towns in New York and northern New Jersey prohibited the Nazis from assembling. By the end of 1934, the Friends of the New Germany had been banned in several municipalities in New Jersey.

Just days after Camp Wille und Macht was discovered, two members of the New Jersey Assembly introduced a bill that was intended to stop the violent breaches of the peace that took place wherever the Nazis gathered and distributed their propaganda. Under the “anti-Nazi” bill, any person who created, distributed, or circulated a “book, speech, article, statement, circular or pamphlet” which “in any way … incites, counsels, promotes, or advocates hatred, violence, or hostility against any group or groups of persons … by reason of race, color, religion, creed, or manner of worship” could be punished with a fine or a jail sentence ranging from ninety days to three years and a fine from $200 to $5,000. The bill did not mention Nazis specifically, but there is no question that it was aimed at the Nazis. The bill banned organizations that had bylaws advocating racial or religious hatred and made it a crime to print, possess, broadcast, or display a photograph or sign that could incite such hatred. It was forbidden under the law to make any statement “in the presence of two or more persons” inciting “race hatred.” Even true statements could be punished if they promoted hatred or hostility against any group of persons in the state by reason of race or religion. The bill was the most sweeping proposal for a hate speech law in American history to that time.

Excerpted from Hate Speech: The Legal and Cultural History of an American Dilemma by Samantha Barbas, courtesy of the University of California Press. Copyright © 2026.

The post The Anti-Nazi Law appeared first on Reason Magazine.

The European Union has placed ChatGPT, Reddit, and Roblox under stricter rules after all three passed the threshold of 45 million average monthly users in the E.U. Under the Digital Services Act (DSA), they must now assess and reduce risks involving illegal content, children’s safety, users’ mental and physical well-being, fundamental rights, elections, and public security. ChatGPT is the first AI chatbot to be subject to the DSA and is being treated as a Very Large Online Search Engine, while Reddit and Roblox are classified as Very Large Online Platforms. The companies will face greater transparency requirements and government oversight. The move is part of the E.U.’s increasing oversight of large tech companies.

The post Brickbat: E.U. Blues appeared first on Reason Magazine.

Earlier tonight, the Supreme Court, in a 7-2 ruling, refused to stay a lower-court decision blocking the Postal Service’s policy implementing Donald Trump’s executive order intended to curb mail-in voting. I have previously written about the First Circuit ruling in this case, and about how Trump’s order is a blatant assault on federalism and separation of powers.

As is often the case with “shadow docket” decisions, the majority opinion offers almost no analysis. But they do state that “[t]he Government is unlikely to succeed on the merits of its challenge to the District Court’s preliminary injunction [against the USPS policy].” That indicates they believe that Trump’s policy is indeed illegal, likely for reasons similar to those outlined by the district court and the First Circuit. 

In a concurring opinion, Justice Brett Kavanaugh states that he thinks “there is at least a fair prospect that the final rule falls within the Postal Service’s statutory authority,” but that the district court was still right to block it, under the Administrative Procedure Act  “because state and local election officials do not have sufficient time to reasonably implement the rule before the elections.” 

In a dissent joined by Justice Thomas, Justice Samuel Alito argues that the policy is within the authority of the USPS, but bases it on statutes that nowhere mention altering mail-in ballot election rules, but only authorize USPS to make rules for the collection and delivery of the mail. Alito largely ignores the massive federalism and separation of powers issues arising from conflating the latter with the former. In addition, as the district court and the First Circuit noted, twisting this language into a sweeping power to override state mail-in ballot rules raises serious “major questions doctrine” issues.

As the District Court noted, Congress must speak clearly when delegating a “highly consequential power” to the executive, and the power to rewrite state election rules is pretty obviously highly consequential. and, to put it mildly, the statutes granting authority to the USPS come nowhere near clearly giving it such power. Indeed, they don’t even mention voting rules.

To put it a different way, inferring a power to regulate mail-in voting from a power to make rules necessary for the delivery of the mail is a classic example what Justice Antonin Scalia famously called “hid[ing] elephants in mouseholes,” warning courts to avoid  assuming Congress did that. The Trump administration claims there is a truly enormous elephant in a very tiny mousehole here. The Court was right to reject that position.

The post Supreme Court Refuses to Block Lower-Court Ruling Against Trump's Mail-in Ballot Executive Order appeared first on Reason Magazine.