One of the disbanded advisory groups was the NIH Center for Scientific Review Advisory Council, created in 1988. That council did not evaluate grant proposals; instead it counseled NIH officials on how to distribute research funds.

“Be outraged”

Outside NIH, the CDC saw nine advisory panels dissolved, including ones like ACIP that were weakened. Four advisory committees were also ended at the FDA.

The report says the abolished or compromised panels covered areas such as childhood vaccines, inherited disorders in newborns and children, Alzheimer’s disease, health equity, infection control in healthcare settings, rural health, novel and exceptional technologies, long COVID, and the US Preventive Services Task Force (USPSTF).

In January, HHS added 21 individuals to the Interagency Autism Coordinating Committee, a panel that tracks autism research and progress on causes, prevention, diagnosis, and treatment. At least eight of those appointees share Kennedy’s belief in the discredited claim that vaccines cause autism. Following the appointments, autism researchers and advocates formed an independent nongovernmental advisory committee to counter the expected misinformation from the federal panel.

At the FDA, the advisory panels that were ended dealt with arthritis, medical imaging drugs, pharmaceutical sciences, and patient engagement.

Overall, the report concludes that Kennedy’s assault on these expert advisory bodies is weakening the biomedical sector and the nation’s health—and that repairing the damage may be slow and difficult.

“All Americans—patients, lawmakers, and scientists alike—have every reason to be outraged by the harm Trump has inflicted on federal health advisory committees,” Michael Abrams, a senior health researcher at Public Citizen and author of the report, said. “Trump’s actions are eroding biomedical research, undermining long-standing processes for approving new drugs and medical devices, and weakening federal vaccine policy. Muzzling and skewing outside experts leaves HHS vulnerable to stagnation and corruption that harms the health of all Americans.”

Google declined to offer any additional comment beyond the blog post it published on its DarkSword findings. WIRED also tried to contact PARS Defense via its X account but did not receive an immediate response.

Lookout says DarkSword is built to exfiltrate information from vulnerable iPhones, including passwords and photos; message logs from iMessage, WhatsApp, and Telegram; browser history; Calendar and Notes entries; and even data from Apple’s Health app. Although the campaign appears to be espionage-focused, DarkSword also harvests cryptocurrency wallet credentials, indicating the operators may have pursued a for-profit cybercrime angle as well.

Rather than installing persistent spyware on victims’ phones, DarkSword employs subtler methods more commonly associated with “fileless” malware on Windows, hijacking legitimate iOS system processes to extract data. “Instead of using a spyware payload to brute force your way through the file system—which leaves tons of artifacts of exploitation that are pretty easy to detect—this just uses system processes the way they’re meant to be used,” iVerify’s Cole says. “And it leaves far fewer traces.”

Cole adds that because of this fileless approach, a DarkSword compromise doesn’t survive a reboot. Instead it grabs data from the device within minutes of the intrusion—a “smash-and-grab” tactic, he says.

While the Coruna iOS hacking toolkit revealed earlier this month targets iOS versions 13 through 17, DarkSword is effective against most builds of iOS 18, the release that preceded last fall’s iOS 26. (DarkSword actually contains two separate exploit “chains” that leverage different vulnerabilities in early and late iOS 18 builds, depending on which one the target device is running.) That means far more phones are vulnerable to DarkSword than to Coruna, particularly given the relatively slow uptake and unpopularity of iOS 26, which has been criticized for new features such as a “liquid glass” interface some users say is overly animated and reduces legibility.

If you’ve spent any time online, you’ve likely used a service like Google Translate to turn webpages or bits of text between languages from Uzbek to Esperanto. But what if you wanted to convert text into more obscure “languages” such as “LinkedIn Speak,” “Gen Z slang,” or “horny Margaret Thatcher”?

This week, lots of people online were amused to discover that the AI-driven Kagi Translate can carry out these and many other unlikely “translation” feats. While the shared discovery highlights the playful, creative side of large language models, it also reveals the hazards of giving users free rein with generalized LLM tools.

What exactly counts as a “language”?

You may know Kagi as the paid rival to Google’s deteriorating search product, but the company rolled out Kagi Translate in 2024, pitching it as a “simply better” alternative to services like Google Translate and DeepL. At launch, Kagi said Kagi Translate “uses a combination of LLMs, selecting and optimizing the best output for each task,” a setup that “can occasionally lead to quirks that we’re actively working to resolve.”

The initial versions of the tool offered simple dropdowns to pick from 244 languages for the source and target. In February 2025, though, at least one little-noticed Hacker News poster found that by tinkering with the URL parameters you could set the target language to “rude man with a Boston accent” without anything breaking.

An HN user spotted the funnier uses of Kagi Translate more than a year ago, but it drew little attention.

In recent weeks, Kagi’s social accounts have shown off the service’s ability to mimic “Reddit Speak” or produce McKinsey consultant jargon with a couple of clicks in Kagi Translate. Early Tuesday, however, those offbeat use cases escaped limited attention after a Hacker News user gleefully reported that “Kagi Translate now supports LinkedIn Speak as an output language.” Further down that busy HN thread, other commenters pointed out that you can change the output language merely by typing into Kagi Translate’s search box, and the underlying AI will try to accommodate you.

Why officials are targeting platforms rather than individual users

The press release said officials “want to introduce a new ban on so-called ‘nudifier’ systems that use AI to create or manipulate images that are sexually explicit or intimate and resemble an identifiable real person without that person’s consent.” It added that “the ban would not apply to AI systems with effective safety measures preventing users from creating such images.”

As Bloomberg reported, the proposed ban would mark a major shift in the EU’s handling of explicit deepfakes, moving beyond simply prosecuting users to holding platforms accountable. Bloomberg said the Grok scandal “epitomized” why regulators needed to change course, noting that “this amendment is the first” EU move “to specifically target AI platforms” that generate and permit distribution of “sexual material without the subject’s consent.”

Though EU officials didn’t name Grok in the press release, regulators had already been examining the AI system while considering what xAI’s controversy might mean for other, less prominent nudify tools. When submitting questions to the European Commission earlier this year, lawmakers warned:

Recent shocking reports of AI-powered nudity applications, such as Grok on X, but also other tools that are freely available online, highlight an increase in AI-driven tools that allow users to generate manipulated intimate images of individuals without their consent, facilitating gender-based cyberviolence and the creation of child sexual abuse material.

Lawmakers urged that “these systems should be banned from the EU market,” stressing that “individual perpetrators”—who “can often be punished under national criminal law”—“are often hard to find.” They argued a better approach would be to act earlier to “prevent widespread image-based sexual violence from the outset.”

With apparent support from Members of Parliament, the amendment’s likely approval is certain to annoy Musk, who is also confronting US lawsuits seeking injunctions over Grok’s nudify outputs. In January, Ashley St. Clair, a mother of one of Musk’s children, became among the first victims to sue. More recently, three young girls in Tennessee filed a proposed class action on behalf of all children allegedly harmed by Grok’s CSAM outputs.

In the EU, public calls for regulatory action are growing as xAI appears unwilling to stop Grok from digitally undressing real people. Michael McNamara, a member of the civil liberties committee, said in the press release that he believes the ban on nudify apps “is something that our citizens expect.”