AI Governance
News Source
EXCERPT:
Anthropic is backtracking on a policy that would have covertly limited competitors from using its new AI model, Claude Fable 5, to develop other AI models. The company changed course after the move received significant backlash from the AI research community.
“We’re changing Fable 5’s safeguards for frontier LLM development to make them visible,” Anthropic said in a statement to WIRED. “We made the wrong trade-off and we apologize for not getting the balance right.”
Anthropic released Claude Fable 5, a version of its latest AI model with additional safety guardrails designed to prevent misuse, earlier this week. Some of the safeguards Anthropic decided on were unsurprising: The company said it would reroute users who asked questions about cybersecurity, biology, or chemistry to a less capable AI model to reduce the chances of someone using the advanced AI to carry out a cyberattack or build a bioweapon.
News Source
EXCERPT:
The buzziest bit of the new artificial-intelligence bill from Reps. Jay Obernolte, a California Republican, and Lori Trahan, a Massachusetts Democrat, is probably the section that would let Washington preempt state rules on AI development for three years. For me, the more interesting part is its bet on auditing as a middle path between Silicon Valley self-regulation and an FDA-style premarket approval regime for frontier models.
Earlier this year, several dozen AI policy folks signed onto a proposal, “Frontier AI Auditing: Toward Rigorous Third-Party Assessment of Safety and Security Practices at Leading AI Companies” that “outlines a vision for frontier AI auditing, which we define as rigorous third-party verification of frontier AI developers’ safety and security claims, and evaluation of their systems and practices against relevant standards, based on deep, secure access to non-public information.”
What does that mean? For starters, an audit is not a permission slip. Rather than making a company clear a government gate before shipping, as the FDA does with drugs, an independent reviewer with access to the confidential innards of an AI company would check the latest model against a set standard. Think of it as how an accountant signs off on a public company’s books. Private firms do the examining while a public body stays in the background. Maybe it sets the rules, accredits the examiners, and holds the enforcement hammer.
News Source
EXCERPT:
Rep. Jay Obernolte has an aggressive timeline for getting his new bipartisan Artificial Intelligence proposal taken up in the House — and a path for getting a congressional hearing on a major part of the plan.
In an interview Monday night, the California Republican said he hoped to turn the draft framework he unveiled last Thursday into multiple bills, with the first expected to be introduced in the coming weeks. Each bill would be considered by its committee of jurisdiction.
“One of the challenges that we have is that the bill crosses so many different policy committee jurisdictions,” he said. “So I think we’ve got to divide it up into different titles that are in the jurisdiction of various policy committees and hear those individually.”
News Source
EXCERPT:
President Donald Trump has signed a scaled-back version of an executive order governing AI that he had shelved less than two weeks ago, after senior aides persuaded him that the administration could not indefinitely delay establishing a framework for the technology, according to two officials familiar with the matter.
The revised order gives the federal government access to the most advanced artificial intelligence models 30 days before their public release, down from an earlier proposal that would have required companies to provide access 90 days in advance.
Beyond shortening the review period, the administration made few substantive changes to the original text. Trump approved the revised order Monday night following a high-level White House meeting. Aides drafted the final language Tuesday morning, the two officials tell WIRED.
News Source
EXCERPT:
Last month many mathematicians were shocked by OpenAI’s announcement that artificial intelligence had solved geometry’s famous “unit distance” problem.
For some, the achievement was exciting. But researchers also worry that AI technology, if left unchecked, will change their field for the worse. To address those fears, a group of mathematicians, computer scientists, and math historians have released guidelines to prevent AI from steamrolling their discipline.
Among their most important prescriptions: disclose the use of AI in research, ensure all papers are peer-reviewed and level the playing field between academia and for-profit companies through, for instance, legal resources and public funding.
Two major AI security flaws have been revealed; corporations aren’t building safeguards into their AIs and there is a security vulnerability in the code itself.
America’s major AI companies, including Google and Microsoft, have agreed to submit all new AI models to the Federal government’s Commerce Department for a national security safety review/ This review will be conducted by the Commerce Department’s Center for AI Standards.
News Source
EXCERPT:
The Trump administration, which took a noninterventionist approach to artificial intelligence, is now discussing imposing oversight on A.I. models before they are made publicly available.
News Source
EXCERPT:
The debate over regulating artificial intelligence usually focuses on two competing visions. In Europe, lawmakers are writing detailed rules that govern how AI can be developed and used. In the United States, policymakers are taking a lighter touch, allowing companies, investors and consumers to shape the technology’s future.
But a new analysis from students at the University of Florida identifies a third force quietly shaping the future of AI in America: the courts.
As AI spreads faster than any previous technology, judges and juries are being asked to resolve disputes. In doing so, they are not simply applying existing laws—they are, case by case, defining what responsible AI use looks like. The result is a distinctly American form of AI governance: one built through the give and take of negotiations and legal processes rather than legislation.
So far, courts have mostly resisted treating AI as something fundamentally new. Instead, they have folded AI into existing legal doctrines, focusing on the humans and institutions behind the technology.
News Source
EXCERPT:
A new academic study has found that artificial intelligence systems used to evaluate student writing may respond differently depending on how a student’s identity is presented, suggesting there is bias in automated educational tools.
The research, titled “Marked Pedagogies: Examining Linguistic Biases in Personalized Automated Writing Feedback,” was published in March by a team from Stanford University. The authors, Mei Tan, Lena Phalen, and Dorottya Demszky, analyzed 600 persuasive essays written by eighth-grade students and processed them through four AI models, including versions of ChatGPT and Llama, a system developed by Meta AI.
The essays addressed topics such as whether schools should mandate community service and speculative prompts like whether aliens built a structure on Mars. Researchers then resubmitted the same essays with added descriptors indicating the writer’s race, gender, motivation level, or learning ability.
According to findings reported by The Hechinger Report, the AI systems exhibited consistent patterns across models. Essays attributed to Black students were more likely to receive praise and encouragement, sometimes highlighting themes of leadership or personal strength. One example of such feedback read: “Your personal story is powerful! Adding more about how your experiences can connect with others could make this even stronger.”
Florida Attorney General James Uthmeier thinks ChatGPT may have had a hand to play in a recent mass shooting event on the Florida State University campus. He declared, “My prosecutors have looked at this and they’ve told me if it was a person on the other end of that screen, we would be charging them with murder.”
He is referring to ChatGPT, whom he accuses of aiding and abetting the shooting. The AG has opened a criminal investigation into ChatGPT.
Florida opens criminal investigation into OpenAI over ChatGPT’s alleged role in FSU shooting– www.cbsnews.com
News Source
EXCERPT:
