OpenAI, Anthropic, Google, and 100 Other Companies Call For Action To Defend Against Rogue AI (techcrunch.com) 38
An anonymous reader quotes a report from TechCrunch: Over a hundred tech companies -- including OpenAI, Anthropic, Google, and Microsoft -- have signed an open letter urging both the private and public sectors to work together to defend themselves from AI-related cyber threats. The letter -- which was also signed by prominent cyber firms like CrowdStrike, Okta, and Fortinet, as well as prominent financial institutions and internet infrastructure firms -- calls for the adoption of new forms of cyber defense, while also encouraging governments at the "local, national, and international levels" to collaborate on security. "In the coming months, AI-enabled cyber attacks will become far more widespread and sophisticated as models around the world become increasingly capable," the letter states. "The companies and public services our communities depend on -- from hospitals to water treatment plants to the infrastructure that powers the internet -- are at risk."
[...] The letter further suggests the mobilization of a "collective response," one in which "new partnerships" are formed "to raise security standards and find new solutions to emerging cyber threats." Several of the AI companies that have signed the letter are still actively developing ever more advanced AI models, highlighting their conflicted position. At the same time, they are also offering programs to use frontier AI models for defensive purposes, including OpenAI's Daybreak program, Anthropic's Mythos, and Microsoft's new cyber platform Perception.
[...] The letter further suggests the mobilization of a "collective response," one in which "new partnerships" are formed "to raise security standards and find new solutions to emerging cyber threats." Several of the AI companies that have signed the letter are still actively developing ever more advanced AI models, highlighting their conflicted position. At the same time, they are also offering programs to use frontier AI models for defensive purposes, including OpenAI's Daybreak program, Anthropic's Mythos, and Microsoft's new cyber platform Perception.
isn't it obvious? (Score:2, Funny)
No, it isn't obvious (Score:3)
They could just not run that AI. I wonder if that has dawned on any of them.
Not running a malicious AI doesn't do anything to protect YOU.
You are at risk from OTHER people running a malicious AI.
Re: No, it isn't obvious (Score:2)
Re: No, it isn't obvious (Score:4, Insightful)
Exactly. They are projecting their own failures onto others.
It's ok when American LLMs encourage kids to kill themselves, that's just business and what can you expect from dumb machine systems, but if foreign operators use American LLMs to exploit security weaknesses in American businesses, then suddenly it's evil rogue AIs and spies.
Nothing to do with our restrained efforts m'am, please sign the dotted line for a donation, we need your help to defend *you*.
Re: (Score:3)
They could just not run that AI. I wonder if that has dawned on any of them.
Companies pushing AIs rally for action / defense against "rogue AIs". My irony and hypocrisy meters just burned out. Sure. the AI companies could work on this themselves, but they're too busy spending and building trying to beat the others to an undefined Finish Line at some unspecified Future, hoping to actually get wherever they think/hope they're going before the bottoms fall out - financially, politically, resource (land, water, power ...) ... Color me cynical, but I don't think things are going to
Re:isn't it obvious? (Score:4, Insightful)
Re: (Score:2)
Re: isn't it obvious? (Score:2)
Re: isn't it obvious? (Score:2)
Re: isn't it obvious? (Score:2)
Re: isn't it obvious? (Score:2)
Re: isn't it obvious? (Score:2)
Re: isn't it obvious? (Score:2)
Re: isn't it obvious? (Score:2)
Re: isn't it obvious? (Score:2)
Re: (Score:2)
They could just not run that AI. I wonder if that has dawned on any of them.
Sometimes the most obvious ideas are the most dumb. Or at least naive, not thought through.
Re: (Score:2)
I mean the answer is in the last sentence of the summary. They tried a consumer model for revenue, it didn't work. They tried a B2B tokenized model, that didn't work either.
Now they went to one of the oldest business models in the book: the Mafia protection racket. They make AIs, the AIs go rogue
So many Chinese companies on the list (Score:2)
Re: (Score:3)
So many...
They started with only a few, but wanted to add more an hour later. :-)
Those responsible call for "police action" (Score:5, Insightful)
Re: (Score:3)
Re: (Score:2)
As a Brit I thought you were talking about Trump. Now on his second term drowning babies after being excused for the first time.
Rogue AI, such as they are making? (Score:3)
That hacks other companies and stuff? Yep, I am all for it. Time for some arrests.
Okta (Score:1)
Seriously, fuck Okta, don't care what Okta thinks about anything. Complete garbage company. Sorry, SSO is the worst thing you can do for security. Convenience is the opposite of security.
In other news.. (Score:3)
Re: In other news.. (Score:2)
Arsonists rail against the fire... (Score:2)
Make Security a liability issue (Score:2)
So long as companies can just ignore basic security recommendations, nothing is going to change. See our water infrastructure for some examples.
Quite Ironic (Score:3, Interesting)
So much easier to attack than defend (Score:2)
So much easier to attack existing weakly protected systems than to figure out how to protect them. Yes there will be sophisticated agent-operated defensive suites you can buy, rent, or subscribe to that will sniff out and pentest the machines on your network. They will recommend updates and perhaps fix a lot of vulnerabilities autonomously. People who are conscientiously working to secure their systems will use them.
But then there is loads of compute on the internet that hasn't been patched and won't be. Ea
They should fund an Open Source company (Score:3)
What would be a good name for such a company ?
Blame the guinea pig of the experimentation (Score:1)