A powerful new AI tool is raising serious security concerns as its own creator calls the technology “terrifying.”
Anthropic, the company behind the popular AI assistant Claude, developed a new model called “Claude Mythos” that is said to be so advanced at writing computer code it has inadvertently become an expert at hacking. The concern? It could pose a major cybersecurity risk if it falls into the wrong hands.
The Capabilities
As reported on Thursday’s AM Update, Anthropic is so worried about what it built it has decided not to release the software publicly out of fear it could be weaponized by bad actors. In a matter of weeks, Mythos discovered more software bugs than many human programmers find in an entire career.
Among the examples: It identified a flaw in OpenBSD, a highly secure operating system often used to protect firewalls and critical infrastructure. The bug could allow someone to crash a system remotely just by connecting to it, and it has been missed for nearly three decades.
Mythos also demonstrated it can combine multiple small bugs inside the core software that powers much of the internet effectively linking them together to move the user from basic access to full control of a machine.
These examples apparently just scratch the surface of what Mythos reportedly can do. It can find – and potentially exploit – software weaknesses across the systems that power billions of devices worldwide, according to the company.
Private Partnerships
Rather than release the technology publicly, Anthropic is partnering with some 40 to 50 tech companies for what is being called “Project Glasswing.” Organizations including Amazon Web Services, Apple, Cisco, Google, Microsoft, and NVIDIA will be able to utilize Claude Mythos to identify bugs within their own systems and strengthen their own security.
“Project Glasswing partners will receive access to Claude Mythos Preview to find and fix vulnerabilities or weaknesses in their foundational systems — systems that represent a very large portion of the world’s shared cyberattack surface,” Anthropic announced in a blog post. “Project Glasswing is an important step toward giving defenders a durable advantage in the coming AI-driven era of cybersecurity.”
Anthropic said it is also in ongoing discussions with the United States government about Claude Mythos’s capabilities.
In an X post earlier this week, Claude creator Boris Cherny did little to quell concerns about the technology. “Mythos is very powerful and should feel terrifying,” he wrote. “I am proud of our approach to responsibly preview it with cyber defenders, rather than generally releasing it into the wild.”
You can get all the day’s headlines by tuning into the AM Update with Megyn Kelly on YouTube, Apple Podcasts, or wherever you like to listen. And don’t forget that you can catch AM Update on SiriusXM’s The Megyn Kelly Channel (channel 111).