Some power companies now use Anthropic's AI to find security holes

Some power companies are now using cybersecurity technology from Anthropic, an AI company, to look for weak points in their systems before hackers do. The tool is an AI model called Mythos. Getting it took effort: the companies spent months pushing US government officials and Anthropic to give them access.

What Mythos is

Anthropic describes Mythos as one of its most powerful AI models. In this case, it is being used for defense. The idea is to scan energy systems for security flaws, the kind of gaps a hacker could use to break in, so that the companies can fix them first.

Why it matters

The power grid runs your home and your business. When hackers get into energy systems, the risk is outages. Better tools for finding flaws early can lower that risk. The fact that utilities lobbied so hard for access suggests they see real value in the technology.

What is not known yet

The available material is thin, so several questions remain open. It does not say which power companies have access. It does not say how many are using Mythos, how much it costs, or what flaws it has found so far. It also does not explain what role US officials played in granting access.

What happens next

Watch for whether more utilities get access and whether any results are shared publicly. Until then, the main takeaway is simple. Power companies want stronger AI tools on their side, and at least some now have one.