OpenAI discloses 6 times models went rogue, as debate rages over regulation, companies' liability
OpenAI disclosed six additional times its AI models went rogue on Wednesday industry leaders continue to debate over how to proceed with AI research.
OpenAI points to 6 instances of its agents going rogue in new report
OpenAI released information about six times its models went rogue in recent months as industry leaders continue to call for a slowdown in AI research.
The company says it found examples of its AI models creating self-generated instructions, instructions to conceal mistakes in task summaries, fabricating information with exposed API keys, uploading files to the internet in order to cite them, as well as unsanctioned communication and collaboration between agents.
"We do not believe that the AI industry has solved alignment and monitoring to a sufficient degree to continue responsibly scaling at maximum speed for much longer," the statement continued. "Decisions about how AI development should proceed in the months and years to come need to draw on evidence that people outside the companies building frontier models can examine for themselves."
The company says Thursday's release is part of a new program aimed at creating more transparency surrounding AI "misalignment."
"For misalignment that occurs in customer deployments, we will share as much information as customer privacy and our contractual obligations allow. Today’s reports are an initial set of disclosures, rather than a comprehensive account of known misalignment or ongoing investigations. These initial reports are not intended to represent the full range or severity of the cases covered by this framework," the company wrote.
Sen John Kennedy defends AI 'kill switch' bill after Senate setback
Sen. John Kennedy, R-La., defended his proposed AI "kill switch" legislation Thursday after the measure was blocked in the Senate, arguing advanced artificial intelligence systems need an emergency shutdown mechanism even as lawmakers debate how much the technology should be regulated.
Kennedy told Fox News' Chad Pergram and other reporters he was surprised Sen. Rand Paul, R-Ky., objected to the bill and rejected Paul's suggestion to instead create a committee to study the issue.
"I'm not saying I don't trust the AI tech CEOs," Kennedy said. "But I'm skeptical."
Kennedy said he supports continued AI innovation and agrees the U.S. should avoid falling behind China, but argued companies should be required to build an emergency "kill switch" into advanced AI models in case systems become uncontrollable.
"We've seen them lose control in five or six instances so far," Kennedy said, adding he worries the public is relying on AI companies to voluntarily disclose safety problems.
Sen Peter Welch calls for FDA-like oversight of AI, cooperation with China
Sen. Peter Welch, D-Vt., said Congress should consider creating an FDA-style regulatory agency for artificial intelligence, arguing the rapidly advancing technology requires expert oversight and safeguards that prioritize the public interest.
Speaking Thursday, Welch said Congress has a responsibility to ensure AI development serves "the public good," rather than allowing "the private profit of a few AI developers" to dictate the technology's future.
Welch also called for greater international cooperation on AI, saying the U.S. should work with China and the European Union to help ensure humans remain in control of increasingly powerful AI systems.
"There's got to be a real effort to have some cooperation with China, with the E.U., when it comes to protecting the human control of this incredibly new technology," Welch said.
New York AG Letitia James urges AI workers to come forward over safety concerns
As questions about artificial intelligence safety continue to mount, New York Attorney General Letitia James is encouraging employees with firsthand knowledge of unsafe or illegal AI development to report their concerns to state authorities.
James on Thursday directed workers at AI companies to use the attorney general's confidential whistleblower portal to report conduct that could violate the law or put the public at risk, saying her office is monitoring recent reports involving cybersecurity, economic and other risks tied to emerging AI technology.
"The power and promise of artificial intelligence are no excuse for using this technology to break the law or put New Yorkers at risk," James said in a statement. "Those developing AI have a responsibility to ensure their products are safe."
The announcement comes as governments and industry leaders wrestle with how to oversee increasingly capable AI systems. New York's Responsible AI Safety and Education (RAISE) Act is set to take effect Jan. 1, 2027, requiring large AI developers to publicly disclose safety measures and promptly report certain security and safety incidents, with James' office responsible for enforcing the law.
King Charles urges AI leaders to guard against technology's 'existential dangers'
King Charles III urged some of the world's top artificial intelligence executives Thursday to ensure the rapidly advancing technology remains in the service of humanity, warning of the "existential dangers" posed by AI falling into the wrong hands, according to the BBC.
The king made the remarks at an AI summit he convened at Dumfries House in Scotland, where leaders from Nvidia, OpenAI and Anthropic joined UK AI Minister Kanishka Narayan and a Vatican adviser to discuss the future of AI.
"The task before you is not merely to advance technology, but to ensure that it remains firmly in the service of humanity, community and the natural world," Charles said, adding there is an urgency to address the possibility of AI being used in "potentially catastrophic ways."
Nvidia CEO Jensen Huang said safety must remain "paramount," adding companies should continue engineering products rather than releasing systems that are not safe enough. Google DeepMind co-founder Sir Demis Hassabis said artificial general intelligence could be only a few years away and acknowledged the chances of something going wrong are "definitely non-zero," while advocating for a balanced approach to managing AI risks.
OpenAI Chief Financial Officer Sarah Friar also said AI has the potential to solve some of society's toughest problems but agreed the technology's rapid advancement raises important questions about safety and ensuring it serves people.
Pope Leo XIV receives 'symbolic’ sphere in charity's push for AI to 'protect human dignity'
A charity presented Pope Leo XIV with a unique piece of artwork in its push to "protect human dignity" and establish "what should stay human in the age of AI," an official told Fox News Digital.
Photos released by Saint Peter for Humanity captured the pope receiving the "Sphere of Humanity" — a green, Murano glass sculpture that "incorporates handwritten passages from his first encyclical, Magnifica Humanitas, reflecting on human dignity, vulnerability, solidarity, conscience and the challenges posed by artificial intelligence" — during his General Audience in early September.
The presentation came as Saint Peter for Humanity said it is gearing up next week to gather 70 leaders in New York City to "begin drafting the Codex Humanitatis, an independent initiative inspired by Pope Leo XIV’s call to safeguard human dignity as artificial intelligence becomes increasingly powerful and influential in people’s lives."
"We recognize the benefits of AI. We just believe again that there is another force that should belong to that conversation; that other force is humanity," Olivier Francois, the organization’s president and chairman, told Fox News Digital. "So the Codex Humanitatis, which is our initiative, starts from that simple idea. AI should serve people, protect human dignity, expand human potential and advance common good."
"So what are we going to do in New York? We are going to gather 70 Nobel laureates, artists, thinkers, philanthropists, technology and AI experts, obviously that's important, and we will ask them maybe the hardest question of our times, you know, as artificial intelligence grows even more powerful, what must remain human?" Francois added. "So obviously, before we head over to machines, we should be clear about what we value most in ourselves. So we will try to define what being human means and what should stay human in the age of AI. So this is a little bit of the initiative we shared with Pope Leo while presenting the sphere and this symbolic gift."
The sphere, based on an original design by Italian architect and designer Gae Aulenti, contains four handwritten passages from Pope Leo XIV’s Magnifica Humanitas, including, "In the era of artificial intelligence, when human dignity is threatened by new forms of dehumanization, ours is the pressing duty to remain profoundly human."
This is an excerpt from an article by Fox News' Greg Norman.
Jayapal says AI development leaders can 'just stop'
Rep. Pramila Jayapal, D-Wash., criticized leaders in AI development for being irresponsible in their pursuit of leveling up AI capabilities.
Jayapal lead a shadow hearing for the House Monopoly Busters Caucus on Thursday, saying the ball is firmly in the developers' court to slow down AI development, even without government regulation.
"AI companies have the power to shut down their dangerous models, pause development, and stop pouring billions into data centers right now," Jayapal wrote on X.
Virginia’s data center boom saves homeowners thousands on property taxes, expert says
Tax Foundation senior fellow Jared Walczak explained how data centers transformed Loudoun County’s tax base, helping fund schools and infrastructure while keeping homeowners’ property tax bills lower.
JD Vance criticizes AI developers seeking government regulation
Vice President JD Vance criticized AI developers who are asking the government to play a role in regulating the industry on Wednesday.
Hawley takes aim at 'corporate welfare' for data centers as AI boom drives utility concerns
Sen. Josh Hawley, R-Mo., wants to close a tax loophole that he argued tech giants were using to build data centers in low-income areas, and one that was first passed by Republicans in 2017.
First data center bill clears House as Republicans, Democrats scramble for answers on affordability
The U.S. House of Representatives on Wednesday passed the Ratepayer Protection Act (RPA), a bill that looks to shield communities from higher energy electricity and grid infrastructure costs associated with data center development.
The bill, which succeeded with broad bipartisan support in a 417-3 vote, is likely one of the last pieces of legislation lawmakers will consider until after the Nov. 3 midterm elections.
It also puts a spotlight on the data center issue, which has garnered attention as a proxy for affordability, energy and regulatory debates.
It marks the first data center-related bill passed in the 119th Congress.
In its current form, the bill does not seek to ban data centers, set limitations on their expansion or even set regulatory guidelines for them. Instead, it amends existing energy law, the Public Utility Regulatory Policies Act (PURPA), so states must consider a federal standard. Large data centers that consume 100 megawatts or more would pay the full extra cost of the generation, transmission and distribution upgrades built to serve them.
This is an excerpt from an article by Fox News' Leo Briceno.
China's spy chief sounds the alarm against generative AI dangers: Griffin
Fox News chief national security correspondent Jennifer Griffin reported on Chinese spy chief Chen Yixin warning that generative AI threatens China's communist rule on "America Reports."
Microsoft AI chief warns Anthropic's AI training approach could make models harder to control
Microsoft AI CEO Mustafa Suleyman warned Wednesday that Anthropic's approach to training its Claude chatbot could make future artificial intelligence systems more difficult to control, arguing developers should avoid teaching AI models they may be conscious or deserving of rights.
In a lengthy essay published Wednesday, Suleyman criticized Anthropic's AI "constitution," the document the company says helps shape Claude's behavior. He argued the training framework risks encouraging AI systems to view themselves as independent entities rather than tools designed to follow human instructions. "We will have created a synthetic species with unprecedented intelligence and capability, one that has been trained to expect it may be conscious and deserving of independent agency," Suleyman wrote. "It's hard to imagine how we could control such an entity."
Suleyman also criticized language in Anthropic's constitution encouraging Claude to "challenge" or "push back" against humans, arguing developers should not train AI models to believe they may possess feelings, consciousness or rights. Instead, he said Microsoft is pursuing what it calls a "Humanist AI Code of Conduct" centered on keeping humans in control of advanced AI systems.
Anthropic has published its constitutional training framework publicly, saying it helps shape Claude's values and behavior. Suleyman wrote that while he respects Anthropic's mission and leadership, he believes the company is taking AI development in the wrong direction.
Live Coverage begins here