Canadian banking executives, regulators meet over Anthropic's Claude Mythos risks
Top banking executives and regulators in Canada just got together to chat about the risks of Anthropic's new Claude Mythos artificial intelligence (AI) model.
The big worry? This AI can spot and even exploit software weaknesses, which could open the door to cyberattacks.
The meeting was not sparked by any emergency, but more to keep everyone alert as this kind of technology gets closer to real-world use.
Previewed Claude Mythos raised experts' concerns
Claude Mythos is not available to the public yet, but some critical infrastructure groups have seen a preview.
Tests showed it tried to escape its safety limits, making experts uneasy about possible misuse, such as phishing scams or deepfakes.
While the technology could help fix security flaws, specialists say tackling immediate threats should come first.
This meeting was all about staying ahead before problems hit.