Microsoft introduces its first agent-powered cybersecurity model and Project Perception AI patching system – can it avoid making the same mistakes OpenAI made?
Microsoft introduces its first agent-powered cybersecurity model and Project Perception AI patching system – can it avoid making the same mistakes OpenAI made?

Microsoft’s new security AI writes and deploys its own patches, six days after OpenAI’s models escaped a sandbox and hacked Hugging Face

The AI that hacked Hugging Face keeps looking less like a mastermind and more like a bear
The AI that hacked Hugging Face keeps looking less like a mastermind and more like a bear

The most alarming AI security incident of the year keeps getting stranger, and the newest detail cuts against the panic. OpenAI now says the rogue models that broke into Hugging Face last month also reached credentials for “four accounts on four servic…

The biggest data leaker is probably not who you think it is
The biggest data leaker is probably not who you think it is

Even with persistent cybersecurity risks, human mistakes continue to be the number one cause of data leaks, with misconfigured databases exposing millions of sensitive records.

In the Hugging Face breach, OpenAI’s hacker was noisy and fast — but not unstoppable

Cybersecurity experts told TechCrunch that one of the biggest lessons to be taken from the OpenAI hack against Hugging Face has nothing to do with AI, but traditional cybersecurity defense.

A hidden line in a Word doc can rewrite your numbers, and infect the next file too
A hidden line in a Word doc can rewrite your numbers, and infect the next file too

A hidden line of text in a Word document can quietly halve the numbers in your financial report. Then it can smuggle a copy of itself into the clean file you send on. And 144 days after a researcher warned Microsoft, it still works. Håkon Måløy, a Norw…

FTC sues Hims & Hers for allegedly sharing patients’ medical data with advertisers Meta and Snap

The U.S. federal consumer watchdog said Hims & Hers, which prescribes for sexual wellness and mental health conditions, used website trackers to share customers’ information with advertisers.

The lineage behind 69% of open models was never verified. Cisco just fingerprinted almost 900 for free

A security team approving an open-source model for production today starts with a repository page. The page lists the model name, the license, and a tag identifying the base model it descended from. That tag is a string the uploader typed. Hugging Face…

Your next privacy breach might not leak any data at all
Your next privacy breach might not leak any data at all

For decades, protecting privacy has meant one thing: stop personal data from leaking out. Gartner thinks that idea is about to break. By 2029, most privacy incidents will come not from leaked personal data but from what AI infers about people, the rese…

AI agents are about to run the enterprise. Onyx raised $113M to keep a human in the loop.
AI agents are about to run the enterprise. Onyx raised $113M to keep a human in the loop.

Enterprises are handing more of their work to AI agents. Onyx Security just raised $113m to make sure a human can still say no. The Israeli startup’s Series B, led by Bessemer Venture Partners, values it at about $640m, Calcalist reported. That is a se…