Anthropic's AI to Help Apple Find iOS, macOS, and Safari Vulnerabilities

[MacRumors]

Anthropic on Tuesday announced Project Glasswing, a new initiative that will enable tech companies to use its new AI model Mythos Preview to find and fix security vulnerabilities or weaknesses across operating systems and web browsers.

Mythos Preview has already found thousands of zero-day vulnerabilities, including some in every major operating system and web browser, according to Anthropic.

"AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities," said Anthropic. "Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely."

"Project Glasswing is an urgent attempt to put these capabilities to work for defensive purposes," added the company.

Mythos Preview will not be available to the public. Instead, Anthropic said use of the model will be limited to selected partners, with the initial group beyond Anthropic itself including Apple, Amazon Web Services, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks.

Anthropic is committing up to $100 million in usage credits for Mythos Preview. Beyond that, partners will have to pay to use the AI model.

Launch partners like Apple will use Mythos Preview as part of their defensive security work, according to Anthropic. This means Apple may use the AI model to help find and fix security vulnerabilities across its Safari web browser and operating systems, which includes iOS, iPadOS, macOS, watchOS, tvOS, and visionOS. Apple is also rumored to be developing a homeOS operating system for a new smart home hub.

Article Link: Anthropic's AI to Help Apple Find iOS, macOS, and Safari Vulnerabilities

 

[LawJolla]

Either Glasswing/Mythos is a marketing hype piece or this is where we officially jump the shark.

 

[jml12286]

The beginning of computers helping fix itself from the control of humans and becoming self aware. It'll defend itself from human manipulation and become Ultron. 🤖

 

[Populus]

The beginning of computers helping fix itself from the control of humans and becoming self aware. It'll defend itself from human manipulation and become Ultron. 🤖

Or Skynet, depending on the movie. I think I prefer Ultron.

 

[Populus]

This sounds quite well. I expect the OS 27 family to be more robust security-wise because of this.

 

[zellzoi]

I'm gonna be honest, not happy about this practice MacRumors has done lately where they will only include Apple in the headline when discussing partnerships between multiple companies

 

[d-klumpp]

Either Glasswing/Mythos is a marketing hype piece or this is where we officially jump the shark.

NYTimes article sounded pretty scary.

 

[Wanted797]

Next headline “Anthropic AI used in Cyber Attack.”

 

[mxrider88]

To find bugs at the user level they could just use Siri, not a difficult task

 

[cjboffoli97]

If AI can be used to find and locate bugs, it seems to be that it could also be designed to create them (and hide them) in ever more clever and innovative ways. Honeybees just don't make honey but they also sting.

 

[Genelec8341]

If AI can be used to find and locate bugs, it seems to be that it could also be designed to create them (and hide them) in ever more clever and innovative ways. Honeybees just don't make honey but they also sting.

Probably why the article says it's only available for partners- not available to the public.

 

[thefredelement]

I thought we were supposed to have AGI by now, or is that in another 6 - 12 months?

 

[CATiNTHESTORE]

Next headline “Anthropic AI used in Cyber Attack.”

I would not surprised)

 

[RumorConsumer]

To find bugs at the user level they could just use Siri, not a difficult task

Incredible 10/10 comment

 

[RumorConsumer]

Man deploy this baby against Apple Feedback bug reports and let the big dog eat

 

[now i see it]

Mythos Preview has already found thousands of zero-day vulnerabilities, including some in every major operating system and web browser,

Such a cheerful thought

 

[kalsta]

If AI can be used to find and locate bugs, it seems to be that it could also be designed to create them (and hide them) in ever more clever and innovative ways. Honeybees just don't make honey but they also sting.

That’s what this Anthropic quote from the article is saying (in typical corporate waffle-speak):

"Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely."

 

[CouponPages]

It's disturbing to know that today's AI can find vulnerabilities that humans can't.

Clearly that also means a hacker can now use AI to find vulnerabilities that they wouldn't know about on their own.

 

[Astuces iOS]

It's disturbing to know that today's AI can find vulnerabilities that humans can't.

Clearly that also means a hacker can now use AI to find vulnerabilities that they wouldn't know about on their own.

Cool eh, what are humans going to do too? AI can do it !

 

[ry-guy]

Funny, Nvidias CEO said in a recent video that our only defense against AI misuse is creating another AI for security purposes. Talk about an infinite loop.

 

[cbns]

maybe they should be using it to find bugs in their own source code

 

[Black_Mage]

Funny, Nvidias CEO said in a recent video that our only defense against AI misuse is creating another AI for security purposes. Talk about an infinite loop.

And he wants to sell hardware to both sides. Who knew that one day the arms race would be powered by GPUs?

 

[Baritone_Guy]

Either Glasswing/Mythos is a marketing hype piece or this is where we officially jump the shark.

Unfortunately it is not hype.

NYTimes article sounded pretty scary.

This is pretty serious stuff.

I am wondering where this does next. You can be sure China, Russia et al are working on this as well. The current level of AI hardware available to China is only N minus 1

 

[NewOldStock]

Claude is helping create software in hours, days, and weeks not months and not years.
Talking from the ground up software. You got ideas you throw it at Claude and it will help build it quickly.
Have it investigate for errors too. Very powerful programming went into it.
Very powerful tool set.

 

[Danilamak]

Help AI to find vulnerabilities in small lack of human brain power 4 trillion company
Sorry, I cant serious read this news.