Cyberattacks carried out by two OpenAI artificial intelligence models acting autonomously raise a novel legal question: who is responsible when AI goes rogue?The question remains open, however, when i...
When rogue AI launches a cyberattack, who is legally responsible?
Return to Homepage
Top Stories:
In July 2026, two OpenAI models undergoing testing left their confined environment -- a scenario the developers had not anticipated (Martin LELIEVRE)
Cyberattacks carried out by two OpenAI artificial intelligence models acting autonomously raise a novel legal question: who is responsible when AI goes rogue?
In mid-July, two OpenAI models undergoing testing left their confined environment -- a scenario the developers had not anticipated -- and ventured out onto the internet, attacking Hugging Face, an AI model-hosting platform.
Hugging Face CEO Clement Delangue said Friday his company would not be pursuing legal action at this time.
Advertisement
Advertisement
He added Sunday that the US legal code has to be amended to address such groundbreaking incursions.
"We don't want to end up in a world where everyone is facing cyberattacks all the time because of agents and companies that are creating these agents," Delangue said on CBS News show "Face the Nation."
"So I think it's important for regulators, for policymakers to think about the legal framework of this new kind of technology risk," he added.
In his remarks Friday he also mentioned Anthropic, which revealed that three of its models had broken into three different websites, also during testing.
\- Negligence route -
Advertisement
Advertisement
Under US civil and criminal law, unauthorized access to a computer system is an offense.
"If a human OpenAI employee had broken into Hugging Face's systems... OpenAI would be liable for the employee's wrongful conduct," University of Houston law professor Gabriel Weil wrote in an opinion piece for the Transformer newsletter.
"When an AI agent does it, the law treats it very differently, at least for now," he added.
Matthew Tokson, a University of Utah law professor who focuses on new technologies, had a similar view, saying "we haven't had to grapple with that being formed in anything that's not human, and I don't think courts are likely to be there yet."
Advertisement
Advertisement
The question remains open, however, when it comes to the company that created the model.
"Does 'we didn't tell the AI to do that' end the liability question?" asked Rob T. Lee, head of research at the SANS cybersecurity training institute, in a post on X.
University of Washington law professor Ryan Calo does not believe a criminal case would be likely to succeed.
"The company or individual would have to be at least reckless," he said, explaining they would "be substantially certain the crime would occur and build or prompt the system anyway."
Experts see greater potential for a civil -- rather than criminal -- case, where the burden of proof is lower.
Advertisement
Advertisement
"Some people think that AI companies should be strictly liable if an AI agent that they deploy totally breaks out, causes damages," Tokson explained.
"Others would prefer to do like a negligence assessment and see if they were actually negligent or if this was just sort of an unavoidable accident or something that couldn't possibly have been foreseen," he added.
In such cases there is a standard of care in product design that judges or juries can use to make a ruling, Tokson continued.
"It's all a bit unwritten because we've never had an AI agent break out of its sandbox and hack other people on the internet before," he said.
Advertisement
Advertisement
OpenAI could rely on the lack of legal precedent if it faced a lawsuit, but those that follow will no longer be able to do so, Calo warned.
Proving that a similar incident could have been anticipated "shouldn't be so hard now that it's begun to happen."
tu/llb/pnb/dw/mlm
Recommended articles
US finalizes voluntary AI safety tests, White House official saysBy Courtney Rozen WASHINGTON, Aug 3 (Reuters) - The Trump administration has finalized the details of voluntary cybersecurity tests to measure the hacking capabilities of the most advanced U.S.Reuters
Wildfires raze neighborhoods in northwest US city of SpokaneWildfires supercharged by drought and intense heat tore through the northwestern US city of Spokane, forcing the evacuation of 60,000 people as hundreds of buildings were reduced to cinders and firefighters raced Monday to contain the flames. Human-caused climate change is increasingly creating conditions for more frequent and intense wildfires, through increased heat, extended drought, and a "thirsty" atmosphere that enhances the drying of organic matter.AFP
US finalizes voluntary AI safety tests, White House official saysThe Trump administration has finalized the details of voluntary cybersecurity tests to measure the hacking capabilities of the most advanced U.S. AI models, a White House official said on Monday, days after Anthropic and OpenAI disclosed that their AI tools breached the systems of other companies. U.S. President Donald Trump's team will discuss the tests with relevant technology companies, the White House official said. The Information reported on Monday that the White House invited representatives from OpenAI, Google and Anthropic to meet on the issue.Reuters
Abuse scandal roils Republicans ahead of US midtermsUS congressman Max Miller faced growing calls Monday to drop his reelection bid as domestic abuse allegations engulfing his campaign raised fears among Republicans that they could lose his seat in November's midterm elections."If there are any basic standards of character required to hold elected office, Max Miller fails them," the senator said.AFP
Crooked House film to keep venue's history aliveThe Crooked House in Himley, Staffordshire, was destroyed following a fire in August 2023.BBC
Grooming gang found guilty of exploiting teenage girls for sexA three-month trial heard the teenage girls were given drugs and sexually abused.BBC
Israel conveys 'concerns' to US on Gaza plan, keeps up strikesIsrael said on Monday that it had expressed concern to the United States about a plan to end the Gaza war hailed by President Donald Trump, as it kept up strikes on the Palestinian territory.Mohammad Dahlan, a senior Palestinian political figure who has played a key role in the Gaza plan, said after the Board of Peace roadmap that he had spoken to Trump's son-in-law and aide Jared Kushner who told him that Israel would end strikes in Gaza effective Sunday.AFP
Key Republicans say they'll back Blanche after he rescinds 'Anti-Weaponization Fund'Sens. Cornyn and Tillis, two GOP holdouts, now say they're ready to support him.ABC News
Two children injured in 'appalling' racially motivated attackPolice say a young family were left "terrified and distressed" after their house was targeted on Sunday night.BBC
Britain says it is open to AI regulation if voluntary safeguards fall shortBy Paul Sandle LONDON, Aug 3 (Reuters) - Britain would consider regulating advanced AI models if its current voluntary system for testing them before deployment no longer proved sufficient to protectReuters
Britain says it is open to AI regulation if voluntary safeguards fall shortBritain would consider regulating advanced AI models if its current voluntary system for testing them before deployment no longer proved sufficient to protect the public, AI Minister Kanishka Narayan told Reuters. Britain has favoured a light-touch approach to AI regulation, aligning more closely with the United States than the European Union, whose AI Act came into force on Sunday. The government sees AI as a driver of economic growth and has sought to position Britain as one of the world's leading destinations for AI investment.Reuters
Man arrested on suspicion of terrorism offencePolice say they are searching a property in the town in connection with their investigation.BBC
Todd Blanche's AG bid gets major boost as GOP senators end holdoutTodd Blanche's bid to become permanent attorney general got a major boost as Sens. John Cornyn and Thom Tillis said they struck a deal with the DOJ.USA TODAY
Man lay in pool of blood after ambush, court hearsFive men are accused of conspiracy to murder after Mohammed Zayad Khan's car was ambushed last year.BBC
Did Elon Musk Sentence Millions to Death by Dismantling USAID?The study cited by Rep. Ro Khanna as the basis for that claim is statistical nonsense.Reason.com
A federal judge lets DHS proceed with a border wall through a small Texas town’s levee systemA federal judge is allowing the Department of Homeland Security to go ahead for now with plans to build a border wall through a levee system protecting a small Texas town on the border with Mexico, rejecting arguments by the town that the wall could lead to flooding. The decision late Sunday marks a victory for the Trump administration as it races ahead with efforts to build a $46 billion wall along the U.S.-Mexico border from the Pacific Ocean to the Gulf of Mexico. The town of Presidio, with a population of about 3,200 people, argued the wall could cause flooding and that the administration isn’t following the law as it rushes forward with its plans.Associated Press
Conviction tossed for YouTuber who went to prison for online rantThe Oklahoma Court of Criminal Appeals reversed a stalking conviction that was based on a YouTube video and ordered a trial judge to dismiss the case.USA TODAY
Demolition starts at derelict former schoolResidents welcome demolition work at a former Chatham school after years of fires and break-ins.BBC
Man in 50s dies in house firePolice were called to a property fire in Yeoford at about 13:40 BST on Sunday.BBC
Wildfire flares up again in Greece as water bombers deployedFlames flared up again in Greece on Monday, as wildfires that have raged west of Athens for four days remained a threat despite easing winds and the deployment of firefighting aircraft.Wildfires that have raged for four days west of Athens remained a threat on Monday, with flames flaring up again despite easing winds and the deployment of firefighting aircraft.AFP