Five Eyes Alliance Warns That New AI Tools Could Supercharge Hacking

Officials from the United States, United Kingdom, Canada, Australia, and New Zealand issued a serious warning on Monday about artificial intelligence and cybersecurity. They said that cutting-edge AI technology could soon make hacking attacks much more powerful and dangerous. The group, known as the 'Five Eyes' intelligence alliance, said that urgent action is needed right now to deal with this growing threat.

The Five Eyes alliance released a three-page statement explaining their concerns. They said that the most advanced AI models are expected to go beyond what the technology industry currently predicts. According to the statement, these tools could completely change both how cyberattacks are carried out and how defenders try to stop them. The officials wrote, 'The timeline is not years, it is months.'

The statement did not go into a lot of specific detail about the threat. Instead, it mostly reminded people of basic cybersecurity advice that experts have recommended for years. For example, officials urged organizations to quickly fix broken or outdated software and to avoid connecting systems to the internet unless it is truly necessary. They also encouraged defenders to use AI as a tool to protect networks, such as by spotting weaknesses faster or responding to attacks more quickly.

The warning points to growing concerns about specific AI models, including one called 'Mythos' made by a company called Anthropic, and another called 'GPT-5.5-Cyber' made by OpenAI. Officials say these tools could allow users to carry out complex and potentially very damaging hacks with much less effort than before. The ease with which these attacks could be launched is what worries intelligence agencies the most.

Earlier in June, the U.S. government ordered Anthropic to suspend access to a version of Mythos for people in other countries. The government said it had concerns about national security, meaning it worried that foreign individuals or groups could use the tool to harm the United States. As a result, Anthropic was forced to shut down that version of the model.

Around the same time, a U.S. agency called CISA — the Cybersecurity and Infrastructure Security Agency — made an important change to its rules. CISA was one of the agencies that signed Monday's warning statement. It shortened the deadline for government officials to fix serious security problems in their computer networks from a longer period down to just three days. Officials said AI threats were the reason for this faster timeline.