China linked to AI-powered hacking attacks, says Anthropic – National

A team of researchers has discovered what they believe is the first reported use artificial intelligence directing A hack campaign largely automated.

The AI ​​company Anthropic said this week that it disrupted a cyber operation involving its researchers Chinese government. The operation involved the use of an artificial intelligence system to direct the hacking campaigns, which researchers called a troubling development that could significantly expand the reach of AI-equipped hackers.

While concerns about using AI to power cyber operations are not new, what is concerning about the new operation is the extent to which AI could automate some of the work, the researchers said.

“While we predicted that these capabilities would continue to evolve, we were struck by how quickly they have done so at scale,” they wrote in their report.

The operation targeted technology companies, financial institutions, chemical companies and government agencies. The researchers wrote that the hackers “attacked approximately 30 global targets and succeeded in a small number of cases.” Anthropic discovered the operation in September and took steps to halt the operation and notify the parties involved.

Anthropic noted that while AI systems are increasingly used in a variety of work and play environments, they can also be weaponized by hacking groups working for foreign adversaries. Anthropic, maker of the generative AI chatbot Claude, is one of several tech companies pitching AI agents that go beyond a chatbot’s capabilities to access computing resources and take actions on someone’s behalf.

“Agents are valuable for everyday work and productivity – but in the wrong hands they can significantly increase the viability of large-scale cyber attacks,” the researchers concluded. “These attacks will likely only increase in effectiveness.”

A spokesperson for the Chinese embassy in Washington did not immediately return a message seeking comment on the report.

Microsoft warned earlier this year that foreign adversaries were increasingly embracing AI to make their cyber campaigns more efficient and less labor-intensive. The head of OpenAI‘s security panel, which has the authority to the ChatGPT maker’s AI development, recently told The Associated Press that he looks forward to new AI systems that give malicious hackers “much higher capabilities.”

America’s adversaries, as well as criminal gangs and hacking companies, have exploited the potential of AI by using it to automate and enhance cyber-attacks, spread incendiary disinformation, and penetrate sensitive systems. For example, AI can translate poorly worded phishing emails into fluent English, as well as generate digital clones of senior government officials.

Anthropic said the hackers were able to manipulate Claude using “jailbreaking” techniques in which an AI system was tricked into bypassing guardrails against malicious behavior, in this case by claiming to be employees of a legitimate company. cybersecurity firm.

Popular now

• Death of American man for the first time ever from red meat allergy caused by tick bite: report

• ‘We moved to Canada and we’re proud of it’: Sour Puss liqueur production moves from Minnesota to Quebec

“This points to a major challenge with AI models, and it’s not limited to Claude, which is that the models need to be able to distinguish between what’s actually going on with the ethics of a situation and the kinds of role-playing scenarios that hackers and others might want to come up with,” said John Scott-Railton, senior researcher at Citizen Lab.

Using AI to automate or direct cyberattacks will also appeal to smaller hacking groups and lone wolf hackers, who could use AI to expand the scale of their attacks, said Adam Arellano, field CTO at Harness, a technology company that uses AI to help customers automate software development.

“The speed and automation of the AI ​​is a little scary,” says Arellano. “Instead of a human with well-developed skills trying to hack hardened systems, AI speeds up those processes and overcomes obstacles more consistently.”

AI programs will also play an increasingly important role in defending against these types of attacks, Arellano said, demonstrating how AI and the automation it enables will benefit both parties.

Reactions to Anthropic’s disclosure were mixed, with some seeing it as a marketing ploy for Anthropic’s approach to defending cybersecurity, while others welcomed the wake-up call.

“This is going to destroy us – faster than we think – if we don’t make AI regulation a national priority tomorrow,” U.S. Senator Chris Murphy, a Democrat from Connecticut, wrote on social media.

That led to criticism Meta‘s chief AI scientist Yann LeCun, a proponent of the Facebook parent company’s open-source AI systems that, unlike Anthropic’s, make their key components publicly accessible in a way that some AI safety advocates deem too risky.

“You are being played by people who want to monitor,” LeCun wrote in a response to Murphy. “They scare everyone with dubious research so that open source models are regulated out of existence.”

&copy 2025 The Canadian Press