FT column warns AI could make cyber attacks easier for more people

In short: A Financial Times columnist argues that as AI gets better, it could make cyber attacks cheaper and easier, even for less skilled criminals.

What's going on

A Financial Times opinion column says AI is starting to reduce the “cost of intelligence”, meaning some kinds of skilled work can be done faster and with fewer people. The author suggests this could change jobs and productivity, but it also has a darker side for online security.

The column points to Anthropic’s claim that its latest AI model, called Claude Mythos, can find weak spots in cyber defenses faster than most humans. A “vulnerability” is like an unlocked window in a building, it is a small weakness someone can use to break in. The author says even if this specific claim is overstated, similar tools are likely coming because AI is already good at writing code, and the same skill can be used to search for flaws.

The piece argues that this could widen the group of people who can launch serious cyber attacks. In the past, some attacks required expert skills. With AI tools, the gap between experts and beginners may shrink, which could raise risks for critical services that rely on computers, such as power grids and water systems.

What to watch

The author predicts an ongoing “arms race” where defenders build stronger protections and attackers use smarter tools to find new weak points. That can mean higher costs, both for businesses and for taxpayers, and more situations where people are asked to prove they are real in person because deepfakes (fake photos, videos, or voices) get more convincing.

Source: Financial Times