UK regulators today ordered Google to put clearer attributions and links to publishers’ content in its AI-generated search features. The UK’s Competition and Markets Authority (CMA) also said Google must give publishers a way to opt out of AI features in search.

“In a world first, publishers will now have effective tools to prevent their content being used to power AI features in search, such as AI Overviews,” the CMA said today. “This will put publishers, like news organizations, in a stronger position to negotiate content deals with Google. To boost consumer trust, Google is also now required to make sure that publisher content is properly attributed, using clear links, in AI‑generated search results.”

The CMA ruled that Google may not penalize publishers for opting out of AI, meaning that Google can’t downrank opted-out publishers in general search results. The CMA said Google will have nine months to comply with all requirements but that the agency “expects important parts of the controls to become available to publishers well before that deadline. Google will also be required to submit and publish compliance reports, supported by key data and metrics, explaining changes it has made and how it has complied.”

Google’s AI Overviews tend to give confident-sounding responses to search queries, but the links to sources in the AI Overviews may or may not support those confident responses. Clearer attribution and links could make it easier for searchers to determine the accuracy of AI Overview summaries.

The CMA applied the rules to Google after determining that it has “strategic market status” in general search services, and has ongoing investigations into Apple and Microsoft. Google today said it will comply with the CMA decision.

https://arstechnica.com/tech-policy/2026/06/google-ordered-to-put-clearer-links-in-ai-search-and-let-uk-publishers-opt-out/

Once covered models are defined, Nguyen then warned that the effectiveness of the safety testing will likely depend on whether AI firms are fully transparent and treat the process as a “genuine collaboration.”

“Underneath the definitional problem sits an observability problem,” Nguyen wrote. “The government cannot assess what it cannot see, and frontier capabilities are visible only to the labs that build them.”

Ferren suggested that “the window for erecting proper cyber defenses to new AI models may also close quickly,” and that even a well-designed government program may struggle to properly vet frontier models in such a short timeframe. “Even when well implemented, pre-deployment testing has limits,” Ferren said, noting that Google’s threat intelligence team has found state-aligned actors using frontier models to automate cyberattacks and “researchers have shown that Mythos-style vulnerability reasoning can be reproduced with open-weight systems.”

So while AI may voluntarily submit to testing, they may be financially motivated to seek a rubber-stamp, rather than work with the government to test known frontier capabilities to their fullest extent.

“It will likely prove difficult to develop models that are incapable of malicious hacking yet remain commercially compelling,” Ferren said.

He concluded that the EO “may yield short-term cybersecurity benefits,” but the “long-term effect” remains “unclear.”

Nguyen suggested the EO takes necessary steps to create “classified cyber benchmarking, voluntary prerelease evaluation, and coordinated vulnerability scanning” that “the national security community will need for decades” to “continuously evaluate systems that are probabilistic rather than deterministic, autonomous rather than directed, and whose capabilities change with every update.”

But the safety testing will have to evolve as fast as the technology does, Nguyen said, otherwise we risk assessing emerging models against “yesterday’s risks.”

That’s why, at its core, the process will depend on an honest exchange between stakeholders with deep technical expertise and confidential national security insights. It’s the only way to ensure the US focuses its energies on protecting the public from the most credible and consequential AI risks, rather than just providing “performative reassurances,” Nguyen wrote.

https://arstechnica.com/tech-policy/2026/06/trumps-ai-executive-order-may-not-prevent-dangerous-deployments/

Amazon declined to comment on the lawsuit when contacted by Ars today.

The Electronic Frontier Foundation (EFF) wrote in November that Ring’s Familiar Faces will scan “many people who have not consented to a face scan, including friends and family, political canvassers, postal workers, delivery drivers, children selling cookies, or maybe even some people passing on the sidewalk.” The EFF said Amazon seems to be “try[ing] to unload some consent requirements onto individual camera owners themselves” with messages reminding customers to comply with applicable laws.

“But Amazon—as a company itself collecting, processing, and storing this biometric data—could have its own consent obligations under numerous laws,” the EFF said, urging regulators to “investigate, protect people’s privacy, and test the strength of their laws.”

Senator urged Amazon to end Familiar Faces

US Sen. Ed Markey (D-Mass.) has urged Amazon to discontinue the Familiar Faces feature. Markey sent Amazon a letter in October 2025 asking how Familiar Faces works, and summarized Amazon’s responses in a February 2026 letter that repeated his call to end Familiar Faces.

Markey said that Amazon revealed in its response to his first letter “that Ring’s privacy protections apply only to device owners who may ‘opt in’ to the Familiar Faces feature, while providing no comparable consent mechanism for individuals unknowingly subjected to facial recognition, leaving members of the public with no right to consent to a facial scan and no control over their biometric data.”

According to Markey’s follow-up letter, Amazon also revealed that “individuals seeking deletion of their biometric data [must] request removal from each individual Ring device owner, forcing people to make separate deletion requests for every home they visit,” and “that the number of law enforcement agencies on its Neighbors Public Safety Service has grown from 2,161 in 2022 to 2,723 today.”

Amazon last year introduced an AI-powered “Search Party” feature advertised as being useful for finding lost pets, which led to backlash after a Super Bowl ad. Amazon subsequently ended a deal with Flock Safety that would have sent Ring customer videos to Flock, which is used by police departments.

Ring posed privacy risks before the Familiar Faces and Search Party features were launched. In 2023, the FTC filed a lawsuit accusing Ring of invading users’ privacy by “allowing thousands of employees and contractors to watch video recordings of customers’ private spaces.” Amazon did not admit any wrongdoing but agreed in a settlement to pay $5.8 million for customer refunds, delete certain types of data, and implement privacy and security controls.

https://arstechnica.com/tech-policy/2026/06/amazon-owned-ring-should-pay-americans-for-scanning-their-faces-lawsuit-says/

By early March, a government program director was telling UCAR that he needed to “get this done quickly” and that documentation of the supercomputing center needed to be handed over “yesterday.” Even now, months after the deadline for public feedback on the decision, the government admits it hasn’t fully evaluated the comments it received. “The sequence of events strongly suggests that the outcome was predetermined,” the decision notes.

For all of those reasons, he concluded that the NSF had already reached a final decision on the transfer of the supercomputing center, and that decision was subject to review under the Administrative Procedures Act, which is what the rest of the case hinged on.

Blocked

As in so many other cases that have made their way into the courts, the government does not seem to have been prepared to offer much of a defense of its actions. The Administrative Procedures Act prohibits actions that are “arbitrary and capricious,” and Jackson found that there was a “failure to articulate any rationale” for the decision to relieve UCAR of its management role.

He noted that some internal documents introduced as evidence indicated that there was dissatisfaction with NCAR’s pursuit of climate research and hosting of scientific programs intended to improve minority participation. But the government chose not to use those as arguments, so the court didn’t need to evaluate them. UCAR, in contrast, introduced significant evidence that the decision to harm NCAR was part of a range of measures meant to pressure Colorado’s Democratic governor about an unrelated matter.

Given that, the court concluded that forcing UCAR to give up its supercomputing center was arbitrary and capricious, and thus violated the Administrative Procedures Act.

UCAR was also able to demonstrate that it was suffering irreparable harm due to the uncertainty about its future. It has experienced unusually high levels of attrition among its staff, who have a rare set of technical skills and require additional training after hiring. And it expects it will be difficult to find replacements for them.

Given those circumstances, Jackson has issued an injunction blocking the government from forcing NCAR or UCAR to give up any resources related to the supercomputing center.

There are still additional threats to NCAR, including breaking it up, transferring other resources, and even selling its Boulder headquarters. So, this victory is far from the end of the threats. But the legal issues that decided the case are likely to apply to the additional threats, unless the government has a defense that it simply chose not to present here.

https://arstechnica.com/science/2026/06/judge-blocks-part-of-trump-admins-effort-to-hurt-colorado-research-center/

Recommendations for humans

So what is a human mathematician to do during the AI boom? The Leiden Declaration recommends that individual mathematicians transparently disclose their use of AI tools, retain responsibility for the correctness of their mathematical work, continue crediting human authors while properly attributing work even if AI tools make that difficult, and consider using only AI tools that align with the values articulated in the declaration

The declaration also reminds mathematicians that mathematics has “applications in the development of technology for use in warfare, oppression, mass surveillance, and the undermining of democracy,” and so mathematicians should make ethical decisions accordingly when choosing external partnerships with tech companies.

Professional mathematical organizations can develop guidelines for the use of AI and other automated tools in publication and review, protect the rights of researchers as authors through licensing agreements that prevent their work from being used as training data without consent, and support the role of peer-reviewed publications. The declaration also suggests such organizations “actively prepare to become involved if major mathematical results are claimed using unconventional means.”

The authors of the declaration also offer straightforward recommendations for policymakers, including “protect the rights of authors,” “regulate the artificial intelligence industry,” and “invest in public computational infrastructure.” Under “don’t believe the hype,” the declaration warns about how “there is currently a strong commercial incentive on the part of the technology industry to overstate the capabilities of their products.”

Lastly, the declaration acknowledges that the tech industry “has offered lucrative jobs, monetary rewards, computing resources, and intellectually stimulating opportunities that some mathematicians have found attractive… in an era of underfunding of higher education and precarious academic employment.” It calls on such collaborations between mathematicians and the tech industry to abide by the standards laid out in the declaration.

“By endorsing the declaration, the IMU affirms that the future of mathematical research must be guided by human judgment, fair and transparent practices, and the shared values of the global mathematical community,” said Ulrike Tillmann, vice president of the International Mathematical Union, in a statement. “Mathematics is, and should always remain, a profoundly human endeavor.”

https://arstechnica.com/tech-policy/2026/06/mathematicians-warn-of-ai-threats-to-profession-as-industry-encroaches/

Altman accused of making ChatGPT unsafe

The lawsuit joins prior suits accusing Altman of callously deploying AI systems without regard for user safety.

In his complaint, Uthmeier recalled how Altman told TED2025 attendees that right now “the stakes are relatively low” for OpenAI to safety-test its products on real users, which he claimed is the only way to iteratively improve them.

“But the stakes aren’t low,” Uthmeier said. “Floridians—including our vulnerable children—have suffered monetary loss, mental health harms, cognitive decline, and physical harm from Defendants’ deceptive, unethical, and recklessly dangerous conduct. Defendants must be held accountable for the harm they have caused and the dangers they and ChatGPT continue to pose to Floridians.”

In a press release, Uthmeier claimed that in rushing products like ChatGPT model 4o to market, OpenAI “ignored internal and external safety warnings, put children at great risk, and allowed a dangerous product to reach millions of Floridians.”

Similarly, the complaint said that Altman must be held “personally liable for the harm he has caused Floridians through his reckless and willful conduct as founder and CEO of OpenAI, including his utter disregard for the risk to human life caused by his firm’s conduct.”

In a loss, OpenAI could face pressure to implement remedies like age-gating free ChatGPT accounts to protect kids, shutting down conversations that discuss violence and suicide, and removing features that the state says deceptively make ChatGPT feel like talking to a human.

Without more parental controls, the state could push for a ban on teens accessing ChatGPT.

“ChatGPT is not safe for teenagers in Florida to use; its use can lead to self-harm, cognitive decline, and behavioral addiction,” the complaint said.

At a press conference live-streamed on X, Uthmeier vowed to work with other states that want to protect kids to hold OpenAI accountable and fielded questions from reporters. When asked if the state planned to pursue all AI companies—not just OpenAI—Uthmeier said Florida is “certainly” looking at other platforms, but “ChatGPT appears to be the most egregious,” with Altman “central” to pushing features that are dangerous to kids.

“Get ready for a fight, and there’s not one more important than this right now,” Uthmeier said.

https://arstechnica.com/tech-policy/2026/06/florida-sues-openai-sam-altman-after-multiple-chatgpt-linked-murders/

Cassel has been working on protecting communities from coal ash pollution for 15 years and said rain and hurricanes amplified by climate change have exacerbated these threats. And those who live near coal ash dumps, she said, continue to discover cancer at a rate that makes them think, “This cannot be normal.”

“EPA, you know the record,” Cassel said. “You made the record.”

Kristina Zierold, a professor at the University of Mississippi, said she has found that children exposed to coal ash are more likely to suffer from depression and have poorer school performance than children who aren’t exposed.

Zierold said she has been researching the health impacts of coal ash on children since 2011 and was awarded a National Institutes of Health grant in 2015 to investigate coal ash and neurobiological health in children 6 to 14 years old.

She and her research team utilized air pollution and dust sampling in the homes of children to collect coal ash and tested children for neurobehavioral and mental health conditions in multiple ways.

If a child performs poorly in school, that can have cascading effects through adulthood, Zierold said. Depression in children can lead to poor social interaction, lack of learning, and in some cases suicide, she said.

“Do you want your children playing on coal ash in parks and playgrounds?” Zierold asked. “Do you want them breathing it in and ingesting it? I don’t.”

Brianna Knisley, the director of public power campaigns at Appalachian Voices, said the 2008 Kingston Fossil Plant coal ash spill was one of the worst industrial disasters in US history. It’s an example of what happens when the EPA leaves coal ash management up to state regulators and utilities, she said.

The 900 workers who cleaned up the spill were denied protective gear and told the coal ash they were working to remove was clean enough to eat. Hundreds of workers became sick and dozens are dead, Knisley said.

Aerial view of the Tennessee Valley Authority’s Cumberland Fossil Plant in Cumberland City, Tenn.

Angie Mummaw, an organizer with Appalachian Voices who lives near the Cumberland Fossil Plant in Tennessee, said she’s tired of communities like hers being treated as sacrifice zones while the coal industry asks for permanent loopholes instead of cleaning up the messes they’ve created.

Knisley has worked with communities where coal ash was used to fill children’s ball fields and seen Tennessee Valley Authority waste piles of the toxic ash piled up behind a public playground, open to the wind. The Tennessee Valley Authority did not immediately respond to questions from Inside Climate News.

“This is coal ash management without strong federal regulation and enforcement,” Knisley said. “States and utilities are not going to keep communities safe.”

This article originally appeared on Inside Climate News, a nonprofit, non-partisan news organization that covers climate, energy, and the environment. Sign up for their newsletter here.

https://arstechnica.com/tech-policy/2026/05/environmentalists-turn-out-in-force-to-oppose-trump-coal-ash-rollbacks/

Grants meet the culture war

The document makes clear what sorts of things might be considered administration priorities and national interest—and they’re largely a war on woke. For example, the Trump administration canceled PEPFAR, a program meant to limit the spread of HIV in Africa; it’s a step that is estimated to lead to hundreds of thousands of deaths. But to the OMB, that’s a good thing, because the alternative was woke: “Far-left activists hijacked the critical work done by the US President’s Emergency Plan for AIDS Relief (PEPFAR), which was established to respond to the AIDS crisis in Africa. Due to wasteful spending, PEPFAR became a left-wing foreign aid entitlement that attempted to promote abortion and gender ideology.”

(Its cited source for that is an editorial from the Heritage Foundation, a far-right-wing think tank.)

While it demands “viewpoint neutral” behavior from everyone receiving money, it has no issues with engaging in viewpoint discrimination itself. For example, it outright bans any funding for “theories of disparate-impact liability,” the idea that apparently race-neutral rules might have impacts that differ based on the race of the people involved. Also banned: any attempts to compensate for the historic discrimination that has kept women and minorities from having equal opportunities in society. That’s considered DEI, and thus forbidden.

Also out: funding for what it terms “gender ideology,” which it defines as an effort to “deny the biological reality of sex or the sex binary in humans.” Apparently, studying human chromosomal disorders, which can result in unusual combinations of X and Y chromosomes, is no longer welcome in the US. “Ending government-sponsored promotion of divisive gender ideology is critical to scientific inquiry, public safety, and trust in government,” the OMB asserts, based on no evidence whatsoever.

There’s also a political litmus test for funding that harkens back to the McCarthy era, when those with “un-American” ideas were ostracized. “OMB proposes a new provision that agencies may consider an applicant’s affiliations with organizations engaged in activities that violate Federal law, undermine public safety or national security, or advocate for the overthrow of the United States Government,” the document notes.

https://arstechnica.com/science/2026/05/the-office-of-management-and-budget-tries-again-to-cripple-us-science/

ABC: FCC order “has no legitimate purpose”

ABC submitted individual filings for WABC-TV in New York; WPVI-TV in Philadelphia; WTVD in Durham, North Carolina; WLS-TV in Chicago; KGO-TV in San Francisco; KFSN-TV in Fresno, California; KTRK-TV in Houston; and KABC-TV in Los Angeles.

The station “submits this license renewal application under protest in response to an unlawful, arbitrary, and unconstitutional Order issued on April 28, 2026, by the Media Bureau,” ABC’s filings said. “The Commission had not demanded early renewal in over five decades. And it has never before demanded simultaneous license renewal applications from a group of stations commonly owned with a network as it has here. The Order has no legitimate purpose.”

ABC said it was filing the applications without waiving any rights and called on the FCC to rescind the order.

“There is no information that the application will reveal that the Commission could not obtain through other means,” ABC wrote. “The Order is inconsistent with a legitimate exercise of investigative authority and is plainly incompatible with the First Amendment. Worse, the Order opens the door to an assault on the Station’s license, while the Commission searches for a legal pretext to achieve its desired goal. This effort to suppress speech under the guise of bureaucratic process must not prevail.”

Carr calls Disney responses “disingenuous”

Carr wrote in an X post yesterday that the “FCC has been investigating Disney for over a year now after reports surfaced alleging that it had been discriminating against people based on race, gender, or other protected characteristics in violation of federal nondiscrimination laws,” and that “Disney only filed these applications to renew their ABC broadcast licenses after the FCC informed the company that their responses to the agency’s investigation had been disingenuous, deficient, and improper.”

ABC said in its filings that the company produced over 11,000 pages of documents in response to a series of FCC requests. ABC said the order to file early license renewals “purports to investigate ‘possible violations’ of the ‘prohibition on unlawful discrimination,’ but never identifies what violation it had in mind.”

https://arstechnica.com/tech-policy/2026/05/trump-fcc-warns-all-broadcasters-to-follow-orders-or-be-punished-like-abc/

Similarly, ICEList.info is designed to act as a sort of wiki, collecting updates on ICE enforcement activity and cataloging detentions, arrests, and deportations. Individual agents are listed “where sufficient evidence exists” to link them to enforcement events, the About page said, and any attempts to post information that could be used to dox agents violate site rules and are deleted.

“False submissions, harassment, or attempts to misuse the platform will be removed,” the About page said.

Dominick Skinner, who owns ICE List, told Wired that “he does not believe that what ICE List does is doxing,” primarily because “ICE List doesn’t post the home addresses of identified agents.”

In a press release, the DOJ said that it considers doxing to be the sharing of “a victim’s Social Security number, home address, home phone number, mobile phone number, and personal email address.” An incomplete Ars review spot checking 100 profiles of ICE agents on ICE List showed only publicly posted professional contact information.

DOJ’s lack of doxing evidence

The Trump administration has routinely relied on bare mentions of threats of doxing to pressure platforms into censoring social media posts showing ICE activity or linking to sites like ICE List, Freedom of Information Act lawsuits have claimed.

But there’s a notable lack of arrests to back up those claims.

As recently as January, the DOJ has insisted that ICE officers are facing an 8,000 percent increase in death threats. But that press release did not specify where that statistic comes from.

Instead, the agency shared the transcript of a single voicemail that was left for an ICE officer in Minnesota on January 24. In it, the caller doesn’t directly threaten violence, but appears try to intimidate the officer by saying that they “hope” that his wife and mom die and that “everything wrong that could go in your life happens.” They tell the officer that they “hope” the officer gets “hit by a bus” and “paralyzed.” And they end by calling the officer a “traitor to the American people” and urging that the officer “should kill yourself.”

https://arstechnica.com/tech-policy/2026/05/doj-sues-states-that-rejected-ice-requests-for-undercover-license-plates/