Emergence of AI Model Sparks Urgent Cybersecurity Dialogues in the U.S.
A newly unveiled AI model, adept at identifying security vulnerabilities across all prominent operating systems and web browsers, has instigated emergency discussions among high-ranking U.S. officials and Wall Street executives.
In light of unparalleled cybersecurity concerns, its developer has opted to curtail public access to the technology.
Anthropic, the company behind the innovation, announced last week its decision to withhold the complete rollout of the Claude Mythos Preview model. The organization cited the considerable risks associated with making such technology broadly accessible.
According to Anthropic, Mythos possesses the capability to uncover and exploit flaws in “every major operating system and every major web browser.” This revelation has led to prompt dialogues with regulators in Washington.
In a bid to address the looming cyber threats associated with this new AI system, Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened an emergency meeting with chief executives from various banks earlier this week.
This high-stakes gathering featured leaders from Citigroup, Morgan Stanley, Bank of America, Wells Fargo, and Goldman Sachs, all of whom are well acquainted with the subject at hand.
Goldman Sachs CEO David Solomon confirmed that his institution has already gained access to the model and is collaborating closely with Anthropic on cybersecurity initiatives. “We’re hyper-aware of the enhanced capabilities of these new models,” Solomon conveyed to analysts on Monday.
- Recommended Products:
- Plaud Note Pro AI Voice Recorder – $189.00
- AIPI-Lite AI Robot – $16.99
- ZNP Digital Badge – $109.99
The UK government’s AI Security Institute has also issued warnings regarding Mythos, asserting that it represents a “step up” from earlier models in terms of potential cyber threats.
Simulations demonstrated that Mythos could execute a 32-step cyber-attack scenario successfully in three out of ten attempts—tasks that would typically demand extensive time investment from human professionals.
Access to Mythos will be restricted to approximately 40 technology companies via an initiative dubbed Project Glasswing.
This consortium, which will include giants such as Amazon, Apple, Microsoft, Google, and JP Morgan Chase, intends to leverage the model for defensive security operations while sharing relevant findings with Anthropic.
In a blog post detailing its decision to limit access, Anthropic underscored that the repercussions for economies, public safety, and national security could be dire.
The company asserted that its AI models have advanced to a stage where they can outstrip even the most competent human experts in detecting and exploiting software vulnerabilities.
In response, OpenAI revealed on Tuesday the launch of GPT-5.4-Cyber, a new model specifically tailored for cybersecurity purposes.
While adopting a more tempered approach than Anthropic’s alarming revelations, the competitor acknowledged that ensuing models will necessitate “more expansive defenses” as their capabilities continue to outpace existing specialized systems.
UK financial regulators are poised to deliberate the risks introduced by Mythos with British banking executives and government officials within the next fortnight through the Cross Market Operational Resilience Group.
This organization includes chief executives alongside representatives from the Treasury, Bank of England, Financial Conduct Authority, and National Cyber Security Centre.
Dario Amodei, co-founder of Anthropic, emphasized during a Monday event hosted by Semafor on the global economy that discussions with the government are ongoing.
“The government has to know about this stuff,” Amodei stated. “Absolutely, we’re talking to them about Mythos, and we’ll talk to them about the next models as well.”
Source link: Technobezz.com.
