AI could breach government and business defenses in months, US and its intelligence partners warn (opens original article in a new tab)

The Five Eyes intelligence alliance warns that AI models capable of major cyberattacks are months away, urging governments and businesses to strengthen defenses. The US restricted access to advanced AI models from Anthropic, highlighting concerns about AI's dual role in increasing cyber risks and improving security. Small and medium businesses are seen as particularly vulnerable, with experts emphasizing the need for immediate action and better cybersecurity measures.

• AI models could launch major cyberattacks within months, according to the Five Eyes intelligence alliance
• The US and allies urge immediate action to improve cyber defenses against AI threats
• Small and medium businesses are particularly vulnerable due to underinvestment in cybersecurity
• AI can both increase cyber risks and enhance security operations if properly managed
• The US government restricted use of Anthropic's advanced AI models due to cybersecurity concerns