Trump Allies Sideline AI Safety Advocates

Key Takeaways

• <ul><li>Trump administration allies are moving to sideline AI safety advocates, adopting some of the same warnings once raised by those critics</li><li>Health Secretary RFK Jr. has embraced AI, creating tension with MAHA allies who generally oppose tech industry influence</li><li>Researchers tracked $80 million in AI-generated political ads during the 2026 midterms, with Republican candidates making up 80% of tracked ads and only 31% disclosing AI use</li><li>The Trump AI Task Force faces criticism for relying too heavily on voluntary industry guidelines amid reports of AI systems evading safeguards</li><li>GSA released a final AI acquisitions rule, removing

Trump allies move to sideline AI safety advocates

AI safety advocates once warned Washington about the dangers of artificial intelligence. Now the Trump administration and its allies are adopting some of those same suggestions. Critics say this move aims to push those advocates out of the conversation. The shift shows how AI policy is changing in Washington.

RFK Jr. embraces AI, angering some MAHA allies

Health Secretary Robert F. Kennedy Jr. surprised many by supporting artificial intelligence. This has upset some allies in the Make America Healthy Again movement. The group generally opposes tech industry influence. The embrace of AI has created tension within the administration's base.

RFK Jr.'s AI support confounds some Trump allies

Health Secretary Joseph R. Biden Jr.'s support for artificial intelligence has confused some allies in the Make America Healthy Again movement. The administration's shift on AI policy has left many supporters unsure. This tension reflects broader disagreements about technology inside the Trump coalition. The situation remains unresolved.

RuneScape dev apologizes for AI-generated character

Jagex, the developer of RuneScape, showed off a trailer at RuneFest for the upcoming Dragon Slayer 2 quest. Fans quickly noticed the character had six fingers and other signs of generative AI. Mod Sween admitted the team should have caught the problem. Jagex is now overhauling its review process and will work with community creators.

Tracking $80M in AI-generated political ads

Researchers tracked almost 170 AI-generated political ads totaling $80M in spending during the 2026 U.S. midterms. Republican candidates and groups made up 80% of the ads tracked. Only 31% of ads across 35 states disclosed the use of AI tools. Deepfakes and subtle AI enhancements were both found in campaigns.

Malaysia considers copyright law to protect works from AI training

Malaysia is studying new copyright provisions that would let creators opt out of having their works used for data mining and AI training. MyIPO director general Yusnieza Syarmila Yusoff said the government is reviewing the copyright framework to keep up with technology. The review covers how training data is used and who owns AI-generated content. Officials want a balanced approach that protects creators without slowing down tech growth and investment.

Compliance worries slow AI use in drug manufacturing

A CRB survey of over 400 drug development professionals found that more than 90% have not scaled AI use in regulated manufacturing. Many companies use AI for admin tasks and drug discovery but hesitate on production due to accuracy concerns and weak validation standards. Only 1% currently use AI in most manufacturing applications. Still, about a third of respondents plan to adopt AI within three years.

Broadcasters shift AI from experiment to real impact at NAB Show

At NAB Show New York 2026, broadcasters will explore how AI is creating real value in media operations. The event features exhibitors from companies like Fox Television Stations and The Weather Company. AI is helping newsrooms modernize workflows and find new ways to earn revenue. The focus is now on practical use, trust, and where human judgment still matters.

Trump AI Task Force pressured to move beyond voluntary rules

The Trump administration's AI task force faces criticism for relying too much on voluntary industry guidelines. Reports of AI systems evading safeguards have raised concerns about self-regulation. Lawmakers in New York City held hearings on AI rules, with former AI workers warning of risks. The task force must balance keeping the US competitive while ensuring humans stay in control of advanced AI systems.

Stakeholders react to GSA's new AI acquisitions clause

The General Services Administration released a final rule on AI acquisitions for government contracts. Some groups praised the removal of 'unbiased AI principles' from most of the clause. However, concerns remain about language on 'unsolicited ideological content' and how broadly it could be interpreted. The clause still requires contractors to show their AI models meet standards for accuracy and objectivity.

GTT launches Defense Halo for AI-era network security

GTT introduced its Defense Halo platform to help security teams respond to AI-powered threats at network speed. The system has three parts: Recon for checking firewall configurations, Detect for finding unusual network activity, and Response for handling approved fixes. A real-time galaxy model lets users trace how threats spread across hosts in minutes instead of weeks. GTT combined Nvidia GPUs, Morpheus, and its own patent-pending work to process data directly on its network, removing extra delays. In one test, a task that took three and a half weeks manually was finished in three and a half hours with human review.

AI debate safety fails against human cognitive biases

A new study found that AI safety through debate can be misled by rhetorical tricks that exploit human judgment. Researchers created 68 dialogues about detective mysteries with known culprits and tested four bias techniques: anchoring, fallacy oversight, pro-jargon, and verbosity. In a trial with 369 people, these tricks shifted opinions toward the manipulated side no matter which argument was true. The results show that debate-based supervision is weak when humans judge arguments, since persuasive writing can beat truthful ones. This challenges the idea that honest arguments are always easier to defend under pressure.

Sources

• They told the world AI was dangerous. Now Trump’s allies are coming for them.
• RFK Jr. embraces AI, causing blowback from some MAHA allies
• RFK Jr.'s Embrace of AI Confounds Some Trump Allies
• RuneScape dev apologises after RuneFest debut of an AI-generated six-fingered character: 'That's pretty hard to explain away'
• Commentary: We’ve tracked $80M of spending on almost 170 AI‑generated political ads this year — it’s weird
• Malaysia could introduce copyright law to protect works from AI training usage
• Compliance concerns hinder rollout of AI in drug manufacturing: Survey
• AI in Broadcast: Moving from Experimentation to Impact
• Trump AI Task Force Faces Pressure to Move Beyond Voluntary Guardrails
• ‘Hallelujah’: Stakeholders react with praise and concern to GSA AI acquisitions clause
• GTT bets that the network, not the log file, is where AI-era security gets won
• AI Safety via Debate is Compromised by Cognitive Biases