In one of those cases, a judge lifted the hold on construction, ruling that a lack of a sound justification for the hold made it “the height of arbitrary and capricious,” a legal standard that determines whether federal decision-making is acceptable under the Administrative Procedures Act. If this were a fictional story, that would be considered foreshadowing.

With no indication of how long the comprehensive assessment would take, 17 states sued to lift the hold on permitting. They were joined by the Alliance for Clean Energy New York, which represents companies that build wind projects or feed their supply chain. Both the plaintiffs and the agencies that were sued asked for summary judgment in the case.

The first issue Judge Saris addressed is standing: Are the states suffering appreciable harm from the suspension of wind projects? She noted that they would receive tax revenue from the projects, that their citizens should see reduced energy costs following their completion, and that the projects were intended to contribute to their climate goals, thus limiting harm to their citizens. At one point, Saris even referred to the government’s attempts to claim the parties lacked standing as “tilting at windmills.”

The government also argued that the suspension wasn’t a final decision—that would come after the review—and thus didn’t fall under the Administrative Procedures Act. But Saris ruled that the decision to suspend all activity pending the rule was the end of a decision-making process and was not being reconsidered by the government, so it qualified.

Because Trump told us to

With those basics out of the way, Saris turned to the meat of the case, which included a consideration of whether the agencies had been involved with any decision-making at all. “The Agency Defendants contend that because they ‘merely followed’ the Wind Memo ‘as the [Wind Memo] itself commands,’ the Wind Order did not constitute a ‘decision’ and therefore no reasoned explanation was required,” her ruling says. She concludes that precedent at the circuit court level blocks this defense, as it would mean that agencies would be exempt from the Administrative Procedures Act whenever the president told them to do anything.

https://arstechnica.com/tech-policy/2025/12/trumps-order-blocking-wind-development-thrown-out-by-court/

OpenAI’s filing came one day after a New York Times investigation revealed how the AI firm came to be involved in so many lawsuits. Speaking with more than 40 current and former OpenAI employees, including executives, safety engineers, researchers, NYT found that OpenAI’s model tweak that made ChatGPT more sycophantic seemed to make the chatbot more likely to help users craft problematic prompts, including those trying to “plan a suicide.”

Eventually, OpenAI rolled back that update, making the chatbot safer. However, as recently as October, the ChatGPT maker seemed to still be prioritizing user engagement over safety, NYT reported, after that tweak caused a dip in engagement. In a memo to OpenAI staff, ChatGPT head Nick Turley “declared a ‘Code Orange,” four employees told NYT, warning that “OpenAI was facing ‘the greatest competitive pressure we’ve ever seen.’” In response, Turley set a goal to increase the number of daily active users by 5 percent by the end of 2025.

Amid user complaints, OpenAI has continually updated its models, but that pattern of tightening safeguards, then seeking ways to increase engagement could continue to get OpenAI in trouble, as lawsuits advance and possibly others drop. NYT “uncovered nearly 50 cases of people having mental health crises during conversations with ChatGPT,” including nine hospitalized and three deaths.

Gretchen Krueger, a former OpenAI employee who worked on policy research, told NYT that early on, she was alarmed by evidence that came before ChatGPT’s release showing that vulnerable users frequently turn to chatbots for help. Later, other researchers found that such troubled users often become “power users.” She noted that “OpenAI’s large language model was not trained to provide therapy” and “sometimes responded with disturbing, detailed guidance,” confirming that she joined other safety experts who left OpenAI due to burnout in 2024.

“Training chatbots to engage with people and keep them coming back presented risks,” Krueger said, suggesting that OpenAI knew that some harm to users “was not only foreseeable, it was foreseen.”

For OpenAI, the scrutiny will likely continue until such reports cease. Although OpenAI officially unveiled an Expert Council on Wellness and AI in October to improve ChatGPT safety testing, there did not appear to be a suicide expert included on the team. That likely concerned suicide prevention experts who warned in a letter updated in September that “proven interventions should directly inform AI safety design,” since “the most acute, life-threatening crises are often temporary—typically resolving within 24–48 hours”—and chatbots could possibly provide more meaningful interventions in that brief window.

If you or someone you know is feeling suicidal or in distress, please call the Suicide Prevention Lifeline number, 1-800-273-TALK (8255), which will put you in touch with a local crisis center.

https://arstechnica.com/tech-policy/2025/11/openai-says-dead-teen-violated-tos-when-he-used-chatgpt-to-plan-suicide/

Meanwhile, the Corvex press release and SEC filings haven’t been changed or corrected. They still say Raymond was already the Corvex CTO and will continue to serve in that role after the merger. The documents make no mention of Bitworks.

Pre-indictment press release

On November 10, Corvex and Movano Health issued their joint press release announcing the merger. Corvex is a private company and Movano a public one, so the transaction requires approval of Movano shareholders. If the merger is completed, the combined company will be public and go by the name Corvex.

The press release says, “Corvex is an AI cloud computing company specializing in GPU-accelerated infrastructure for AI workloads. Corvex is based in Arlington, Virginia, and is led by Seth Demsey and Jay Crystal, Co-Chief Executive Officers and Co-Founders, and Brian Raymond, Chief Technology Officer.” It goes on to say that after the merger, the combined company will be led by Demsey, Crystal, Raymond, “and other members of the Corvex management team.”

The “is led by” phrase in the press release clearly indicates that Raymond was already the CTO, while the additional statement about the post-merger company indicated he would continue as CTO after the merger’s completion. At the same time, Raymond announced on LinkedIn that he had “formally joined Corvex as the CTO, driving AI at scale for customers around the world.”

The Corvex/Movano joint press release naming Raymond as CTO was submitted to the SEC as an exhibit to a Movano filing about the Corvex/Movano merger. A merger agreement submitted to the SEC by Corvex and Movano includes another exhibit listing three “post-closing officers,” specifically Demsey, Crystal, and Raymond.

The timing of Corvex’s statements about Raymond being its CTO could hardly have been worse. Raymond was indicted in a federal court on November 13 and the indictment was unsealed last week. The US Justice Department alleged that Raymond operated an Alabama-based electronics company through which he supplied Nvidia GPUs to his alleged conspirators “for illegal export to the PRC [People’s Republic of China] as part of the conspiracy.”

https://arstechnica.com/tech-policy/2025/11/tech-firms-new-cto-gets-indicted-company-then-claims-he-was-never-cto/

During the first six months of Trump’s term, 154,000 federal employees signed up for the deferred resignation program, Reuters reported, while more than 70,000 retired. Both numbers were clear increases (tens of thousands) over exits from government in prior years, Kamarck’s report noted.

“A lot of people said, ‘the hell with this’ and left,” Kamarck told Ars.

Kamarck told Ars that her report makes it obvious that DOGE “cut muscle, not fat,” because “they didn’t really know what they were doing.”

As a result, agencies are now scrambling to assess the damage and rehire lost talent. However, her report documented that agencies aligned with Trump’s policies appear to have an easier time getting new hires approved, despite Kupor telling Reuters that the government-wide hiring freeze is “over.” As of mid-November 2025, “of the over 73,000 posted jobs, a candidate was selected for only about 14,400 of them,” Kamarck reported, noting that it was impossible to confirm how many selected candidates have officially started working.

“Agencies are having to do a lot of reassessments in terms of what happened,” Kamarck told Ars, concluding that DOGE “was basically a disaster.”

A decentralized DOGE may be more powerful

“DOGE is not dead,” though, Kamarck said, noting that “the cutting effort is definitely” continuing under the Office of Management and Budget, which “has a lot more power than DOGE ever had.”

However, the termination of DOGE does mean that “the way it operated is dead,” and that will likely come as a relief to government workers who expected DOGE to continue slashing agencies through July 2026 at least, if not beyond.

Many government workers are still fighting terminations, as court cases drag on, and even Kamarck has given up on tracking due to inconsistencies in outcomes.

https://arstechnica.com/tech-policy/2025/11/doge-doesnt-exist-anymore-but-expert-says-its-still-not-dead/

Citing export controls that took effect in 2022, the indictment said the US is trying to disrupt China’s plan to build exascale supercomputers for military and surveillance use. “These capabilities are being used by the PRC for its military modernization efforts and in connection with the PRC’s weapons design and testing, including for weapons of mass destruction, as well as in connection with the PRC’s development and deployment of advanced AI surveillance tools,” the indictment said.

The Justice Department said the conspirators used Janford Realtor, LLC, a Florida-based company that was not involved in real estate despite its name, “as a front to purchase and then illegally export controlled GPUs to the PRC.” Ho and Li owned and controlled Janford Realtor, while Raymond operated an Alabama-based electronics company that “supplied Nvidia GPUs to Ho and others for illegal export to the PRC,” the Justice Department said.

Kickbacks, money laundering

The conspirators paid each other “kickbacks” or commissions on the sale and export of the Nvidia chips, the indictment said. The money laundering charges involve a variety of transfers from two Chinese companies to Janford Realtor and the Alabama electronics company, the indictment said. The indictment lists nine wire transfers in amounts ranging from $237,248 to $1,150,000.

Raymond was reportedly released on bond, while the other three alleged conspirators are being detained. “This is an extremely serious offense. At the time these were being exported, these were Nvidia’s most advanced chips,” US prosecutor Noah Stern told a magistrate judge in Oakland yesterday, according to Wired.

Stein also said in court that “text messages obtained by authorities show Li boasting about how his father ‘had engaged in similar business on behalf of the Chinese Communist Party,’” Wired reported. Stern said that in the messages, Li “explained that his father had ways to import” the Nvidia chips despite US export controls.

https://arstechnica.com/tech-policy/2025/11/tech-company-cto-and-others-indicted-for-exporting-nvidia-chips-to-china/

“I hope to see the Supreme Court rule swiftly to provide businesses the certainty they need,” Shapiro said, arguing in a second post that tariffs “cause uncertainty for businesses, snarl supply chains, and drive inflation and higher costs for consumers.”

As tech companies wait to see how the court rules and how Trump responds to the conclusion of the Commerce Department’s probe, uncertainty remains. CTA’s vice president of international trade, Ed Brzytwa, told Ars that the CTA has advised tech firms to keep their receipts and document all tariff payments.

How chip tariffs could raise prices

Without specifying what was incorrect, a White House official disputed Reuters’ reporting that Trump may shift the timeline for announcing semiconductor tariffs, saying simply “that is not true.”

A Commerce Department official said there was “no change” to report, insisting that the “administration remains committed to reshoring manufacturing that’s critical to our national and economic security.”

But neither official shared any details on when tariffs might be finalized, Reuters reported. And the Commerce Department did not respond to Ars’ request for information on when the public could expect to review findings of its probe.

In comments submitted to the Commerce Department, the Semiconductor Industry Association warned that “for every dollar that a semiconductor chip increases in price, products with embedded semiconductors will have to raise their sales price by $3 to maintain their previous margins.” That makes it easy to see how semiconductor tariffs risk significantly raising prices on any product containing a chip, depending how high the tariff rate is, including products like refrigerators, cars, video game consoles, coffee makers, smartphones, and the list goes on.

It’s estimated that chip tariffs could cost the semiconductor industry more than $1 billion. However, the bigger threat to the semiconductor industry would be if the higher prices of US-made chip made it harder to compete with “companies who sell comparable chips at a lower price globally,” SIA reported. Additionally, “higher input costs from tariffs” could also “force domestic companies to divert funds away from R&D,” the group noted. US firms that Trump wants to promote could rapidly lose their edge in such a scenario.

https://arstechnica.com/tech-policy/2025/11/keep-your-receipts-tech-firms-told-to-prepare-for-possible-tariff-refunds/

The FTC chairman would be required to issue a policy statement detailing “circumstances under which State laws that require alterations to the truthful outputs of AI models are preempted by the FTC Act’s prohibition on engaging in deceptive acts or practices affecting commerce.”

When Cruz proposed a moratorium restricting state AI regulation in mid-2025, Sen. Marsha Blackburn (R-Tenn.) helped lead the fight against it. “Until Congress passes federally preemptive legislation like the Kids Online Safety Act and an online privacy framework, we can’t block states from making laws that protect their citizens,” Blackburn said at the time.

Sen. Maria Cantwell (D-Wash.) also spoke out against the Cruz plan, saying it would preempt “good state consumer protection laws” related to robocalls, deepfakes, and autonomous vehicles.

Trump wants Congress to preempt state laws

Besides reviving the Cruz plan, Trump’s draft executive order seeks new legislation to preempt state laws. The order would direct Trump administration officials to “jointly prepare for my review a legislative recommendation establishing a uniform Federal regulatory framework for AI that preempts State AI laws that conflict with the policy set forth in this order.”

House Majority Leader Steve Scalise (R-La.) this week said a ban on state AI laws could be included in the National Defense Authorization Act (NDAA). Democrats are trying to keep the ban out of the bill.

“We have to allow states to take the lead because we’re not able to, so far in Washington, come up with appropriate legislation,” Sen. Jack Reed (D-R.I.), the ranking member on the Armed Services Committee, told Semafor.

In a Truth Social post on Tuesday, Trump claimed that states are “trying to embed DEI ideology into AI models.” Trump wrote, “We MUST have one Federal Standard instead of a patchwork of 50 State Regulatory Regimes. If we don’t, then China will easily catch us in the AI race. Put it in the NDAA, or pass a separate Bill, and nobody will ever be able to compete with America.”

https://arstechnica.com/tech-policy/2025/11/trump-revives-unpopular-ted-cruz-plan-to-punish-states-that-impose-ai-laws/

US Rep. Richard Hudson (R-N.C.), the subcommittee chairman, defended the bills at today’s hearing. “These reforms will add much-needed certainty, predictability, and accountability to the broadband permitting process and help expedite deployment,” he said.

Cable lobby group NCTA called the hearing “important progress” toward “the removal of regulatory impediments that slow deployment to unserved areas.” Another cable lobby group, America’s Communications Association, said the permitting reform bills “will strip away red tape and enable broadband, cable, and telecommunications providers to redirect resources to upgrading and expanding their networks and services, especially in rural areas.”

$42 billion program delays

Much of the debate centered on a $42 billion federal program that was created in a November 2021 law to subsidize broadband construction in areas without modern access. The Trump administration threw out a Biden-era plan for distributing the Broadband Equity, Access, and Deployment (BEAD) program funds, forcing state governments to rewrite their plans and cut costs, delaying the projects’ start. Money still hasn’t been distributed, though the Trump administration today said it approved the rewritten plans of 18 states and territories.

Hudson alleged that BEAD suffered from “four years of delays caused by the Biden-Harris administration,” though the Biden administration had approximately three years to set up the program. Hudson said that “permitting reform is essential” to prevent the money from being “tied up in further unnecessary reviews and bureaucratic delays.”

The bills set varying deadlines for different types of network projects, ranging from 60 days to 150 days. One bill demands that permit fees for BEAD construction projects be based on the local government’s “actual and direct costs.” Another stipulates that certain environmental and historical preservation reviews aren’t required when removing equipment targeted by a 2019 law on foreign technology deemed to be a security risk.

Rep. Doris Matsui (D-Calif.), the subcommittee’s top Democrat, said during the hearing that she won’t support “proposals that force local governments to meet tight deadlines without any extra staff or funding.” She said that if the “shot clock” specified in the legislation “runs out, the project is automatically approved. That may sound like a way to speed things up but in reality, it cuts out community input, leads to mistakes and sets us up for more delays down the road. If we want faster reviews, we should give local communities more help, not take away their say.”

https://arstechnica.com/tech-policy/2025/11/gop-overhaul-of-broadband-permit-laws-cities-hate-it-cable-companies-love-it/

“The only justice I’m getting is exposing this man and having posted that video,” she added. “And it has gone viral. Now he can live with shame and embarrassment if people have seen it.”

“I’m the victim!” she said. “Is this making sense to any-fucking-body?”

Her numerous videos attracted huge followings—anywhere from 5 million to 30 million views each—and DoorDash eventually felt the need to respond.

“DoorDash never deactivates someone for reporting [sexual assault]—full stop,” said the company.

But, it added, “posting a video of a customer in their home, and disclosing their personal details publicly, is a clear violation of our policies. That is the sole reason that this Dasher’s account was deactivated, along with the customer’s, while we investigated. We’ve also ensured that the Dasher has full access to their earnings.”

Meanwhile, the police were doing something—but not something that Henderson wanted.

The cops determined that the nude man in question “was incapacitated and unconscious on his couch due to alcohol consumption.” Being drunk and naked inside your own home apparently does not qualify as sexual assault on a delivery driver, and the police department said in a press release yesterday that “the investigation by the Oswego Police Department determined that no sexual assault occurred.”

As part of their investigation, the cops found that Henderson had filmed the man and “subsequently posted the video to social media, where it drew significant attention.” This shifted their attention to Henderson’s decision to film and upload the video without the man’s consent.

The police eventually arrested Henderson, who is now charged with two felonies: “Unlawful Surveillance in the Second Degree” and “Dissemination of an Unlawful Surveillance Image in the First Degree.” She was released after being charged, and her case will be heard by the Oswego City Court.

Henderson has stopped releasing videos on TikTok about the situation.

https://arstechnica.com/tech-policy/2025/11/doordasher-claims-sexual-assault-but-cops-arrest-her-for-filming-nude-drunk-man/

Other factors stemmed from social changes, the judge suggested, describing the fourth factor as a trend where Meta app users started feeling “increasingly bored by their friends’ posts.”

“Longtime users’ friend lists” start fresh, but over time, they “become an often-outdated archive of people they once knew: a casual friend from college, a long-ago friend from summer camp, some guy they met at a party once,” Boasberg wrote. “Posts from friends have therefore grown less interesting.”

Then came TikTok, the fifth factor, Boasberg said, which forced Meta to “evolve” Facebook and Instagram by adding Reels.

And finally, “those five changes both caused and were reinforced by a change in social norms, which evolved to discourage public posting,” Boasberg wrote. “People have increasingly become less interested in blasting out public posts that hundreds of others can see.”

As a result of these tech advancements and social trends, Boasberg said, “Facebook, Instagram, TikTok, and YouTube have thus evolved to have nearly identical main features.” That reality undermined the FTC’s claims that users preferred Facebook and Instagram before Meta shifted its focus away from friends-and-family content.

“The Court simply does not find it credible that users would prefer the Facebook and Instagram apps that existed ten years ago to the versions that exist today,” Boasberg wrote.

Meta apps have not deteriorated, judge ruled

Boasberg repeatedly emphasized that the FTC failed to prove that Meta has a monopoly “now,” either actively or imminently causing harms.

The FTC tried to win by claiming that “Meta has degraded its apps’ quality by increasing their ad load, that falling user sentiment shows that the apps have deteriorated and that Meta has sabotaged its apps by underinvesting in friend sharing,” Boasberg noted.

https://arstechnica.com/tech-policy/2025/11/meta-wins-monopoly-trial-convinces-judge-that-social-networking-is-dead/