AI models capable of cyberattacks months away, says Five Eyes
What's the story
A coalition of international intelligence agencies has warned that advanced artificial intelligence (AI) models capable of launching major cyberattacks could be just months away. The warning was issued by the Five Eyes alliance, which includes the United States, United Kingdom, Canada, Australia and New Zealand. The group urged governments and businesses to "act now" to bolster their defenses against these sophisticated threats.
Technological evolution
Transformation in cyber capabilities
The Five Eyes alliance said in a statement, "Frontier AI models are anticipated to exceed current industry expectations, fundamentally transforming both offensive and defensive cyber capabilities." They added that this transformation would happen in months, not years. The group highlighted the rapid evolution of artificial intelligence (AI) and its potential to drastically change cyber risks.
Cybersecurity concerns
Worrying implications for global businesses
The Five Eyes leaders also expressed their concerns over the advanced capabilities of AI technology. They said it lowers "barriers for malicious actors and increases the speed and complexity of attacks." AI experts said the message is "really stark" with worrying implications not just for governments and corporations but also small and medium businesses globally.
Regulatory measures
US orders halt on AI model usage by foreign nationals
In response to the advanced capabilities of an AI model, the Donald Trump administration has ordered AI giant Anthropic to suspend use of its most advanced models by foreign nationals. This broad directive was one of the farthest-reaching actions taken by a government in response to these concerns. The move comes after widespread cybersecurity fears were raised over Mythos, a model that Anthropic claimed was extremely adept at finding security flaws.
Countermeasures
Five Eyes suggest investing in cyber defenses
The Five Eyes leaders have also suggested that businesses and leaders invest in cyber defenses, upgrade old systems or patch faulty software, and limit access to critical systems. They emphasized that while AI is being used by adversaries to "move faster and more effectively," it can also be part of the solution. The security alliance said organizations integrating AI tools into their security operations can detect vulnerabilities earlier, improve software quality, monitor unusual behavior, and respond faster to incidents.