A team of researchers has unveiled what they say is the first use of artificial intelligence to automate a hacking campaign.
AI firm Anthropic said this week that it had disrupted cyber operations that researchers linked to the Chinese government. The operation involved the use of an artificial intelligence system to guide hacking campaigns, which researchers called a disruptive development that could greatly increase the reach of AI hackers.
While concerns about the use of AI to drive cyber operations are not new, what concerns the new operation is the new way in which AI has been able to perform a task, the researchers said.
“While we predict that these skills will continue to emerge, our experience is how quickly they have done so,” they wrote in their report.
Targeted companies, financial institutions, chemical companies and government agencies. The investigators wrote that the hackers attacked “restrictions all over the world and succeeded in a small number of cases.” Anthropic discovered the work in September and took steps to shut it down and notify affected parties.
Anthropic noted that while AI systems are increasingly being used in various forms of work and leisure, they can be augmented with tools by hacking teams working for adversaries abroad. Anthropic, Maker of Ai Chaude Claude, is one of the many technology companies onboarding AI “Agents” that go beyond the power of conversation to discover computer tools and take actions on Quepl.
Get news from around the country
Get the top stories of the Day, politics, economics, and current affairs, delivered to your inbox once a day.
“Agents are useful for daily work and productivity – but in the wrong hands, they can greatly increase the effectiveness of large-scale cyberattacks,” the researchers concluded. “These attacks are only likely to increase with their performance.”
A spokesman 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 using AI to make their campaigns more efficient and effective. The head of the security panel at Openai, which has the authority to stop the development of the chatgpt AI maker, recently told the media that they are looking for new AI programs with “very high capabilities.”
America’s enemies, as well as America’s criminals and hacking companies, have exploited the ability of AI, using it to automate and develop cyberattacks, broadcast distrattacks, deploy dilvalforation discforms and penetrate sensitive services. AI can translate negative depishing emails into English, for example, and generate digital clones of high-ranking government officials.
Anthropic said that the hackers were able to manipulate the length of the strategy, using “sarcastic” techniques that combine its behavior with malicious behavior, this time by claiming to be employees of a legitimate cyberseri company.
“This points to a major challenge with AI models, and it is not limited to Claude, which is that the models must be able to distinguish between the behavior of a misleading situation and others who want to cook,” said John Scott-Railton, a senior researcher in the elements of citizens.
The use of AI to exchange or direct cyberattacks will also attract small hacking groups and wolf hackers, who can use AI Arellano, a technology company that uses AI to help clients improve software development.
“The speed and automation provided by AI is what makes it less scary,” Arellano said. “Instead of a person with prestigious skills trying to hack into complex organizations, AI accelerates those processes and gets past obstacles.”
AI systems will also play a very important role in protecting against these types of attacks, said Arellano, showing how they go and allow that they will benefit both parties.
Reaction to the anthropic revelation was mixed, with some seeing it as a ploy to advertise the anthropic approach to cybersecurity and others accepting its wake-up call.
“This will destroy us – sooner than we think – if we do not make the control of AI a national priority,” said Sur Sen. Chris Murphy, Conficut Democrat, on social media.
That led to criticism from the anchor of Meta ai Scientist Yann Lecun, unlike the anthropic open company, unlike their important elements are publicly available in a way that some advocates of AI security See more.
“He’s being played by people who want to be held in control,” Lecun said in response to Murphy. “They intimidate everyone with questionable studies so that open source models are regulated for the sake of existence.”
& Copy 2025 Canadian machine
