\n But the rise of AI agents, or AI assistants that can carry out tasks autonomously, takes that risk to another level, some experts warn. A single AI agent could scan for vulnerabilities and potentially take advantage of them faster and more persistently than hundreds of human hackers.\n <\/p>\n
\n \u201cThe agentic attackers are coming,\u201d said Shlomo Kramer, founder and CEO of cybersecurity and networking company Cato Networks. \u201cThis is a watershed event in the history of cybersecurity.\u201d\n <\/p>\n
\n Details about Mythos leaked in an unpublished blog post first reported on by Fortune. Anthropic did not respond to CNN\u2019s request for comment. But the company told Fortune<\/a> the leak was a result of human error within its content management system.\n <\/p>\n \n \u201cAlthough Mythos is currently far ahead of any other AI model in cyber capabilities, it presages an upcoming wave of models that can exploit vulnerabilities in ways that far outpace the efforts of defenders,\u201d Anthropic said in the draft<\/a>.\n <\/p>\n \n The company is letting certain organizations test the model ahead of time to improve their systems \u201cagainst the impending wave of AI-driven exploits,\u201d it said.\n <\/p>\n \n Anthropic is also privately warning government officials about the potential for large-scale cyberattacks enabled by Mythos, according to Axios<\/a>.\n <\/p>\n \n But every lab\u2019s next model will pose increasingly severe cybersecurity threats, Kramer told CNN.\n <\/p>\n \n \u201cBehind Mythos is the next OpenAI model, and the next Google Gemini, and a few months behind them are the open-source Chinese models,\u201d he said.\n <\/p>\n \n AI is making it possible to exploit vulnerabilities almost immediately after discovering them, said Evan Pe\u00f1a, chief offensive security officer at cybersecurity firm Armadin.\n <\/p>\n \n But there are still limits to what the models can do, according to Pe\u00f1a.\n <\/p>\n \n Advanced AI models are good for researching software vulnerabilities and developing code to exploit them. But they lack the context a human hacker would have on what an organization\u2019s most valuable information to steal is, Pe\u00f1a said.\n <\/p>\n \n There will always be room for humans in a cyberattack using AI, said Joe Lin, said Joe Lin, co-founder and CEO of Twenty, a firm that sells offensive cyber capabilities to the US government.\n <\/p>\n \n \u201cWe must ensure we are building weapons systems where humans remain firmly in control of decisions and outcomes, because while the machine handles the execution, the human must always own the consequences,\u201d he said.\n <\/p>\n \n An example of how AI has made relatively unskilled hackers more dangerous came in January, when a Russian-speaking cybercriminal used multiple AI tools to hack over 600 devices running a popular firewall software in more than 55 countries, according<\/a> to Amazon Web Services\u2019 security research team. The hacker used generative AI services to \u201cimplement and scale well-known attack techniques throughout every phase of their operations, despite their limited technical capabilities,\u201d AWS said<\/a>.\n <\/p>\n \n The hacker used Anthropic\u2019s Claude model as well as Chinese-made DeepSeek in the attack, according to Eyal Sela, director of threat intelligence at Gambit Security. At one point, the hacker asked Claude in Russian to create a web panel for managing hundreds of the hackers\u2019 targets, according to chat logs the hacker had with AI models that Sela shared with CNN.\n <\/p>\n \n AI gives hackers of varying skill \u201csuperpowers\u201d by simplifying the technical knowledge required to exploit systems, according to Sela.\n <\/p>\n![\"People](\"https:\/\/www.newsbeep.com\/ca\/wp-content\/uploads\/2026\/04\/2026-03-22t015825z-653148237-rc289kauncb7-rtrmadp-3-usa-ai-protests.JPG\")
<\/p>\n