Goldman Sachs Sounds Alarm on Anthropic's AI Model 'Mythos'

Goldman Sachs is voicing serious concerns about the cybersecurity implications of Anthropic’s new AI model, dubbed ‘Mythos.’ According to The Cyber Express, Goldman’s CEO David Solomon stated the financial giant is “hyperaware” of the risks posed by advanced AI, including large language models and Mythos, which can identify and exploit software vulnerabilities at an unprecedented level.

This sophisticated capability has sent ripples through the financial sector. The Cyber Express reports that Mythos is designed to autonomously discover and exploit vulnerabilities, potentially lowering the barrier for cyberattacks and enabling less technically skilled actors to target systems. Anthropic itself has acknowledged the model’s prowess, restricting its access and opting for a controlled preview program called ‘Project Glasswing’ with select organizations like JPMorgan, Apple, Google, Microsoft, Nvidia, and Goldman Sachs.

The initiative aims to bolster defenses before any wider rollout. The Cyber Express notes that Goldman Sachs is actively collaborating with Anthropic and other cybersecurity partners to comprehend and mitigate these emerging threats, emphasizing their long-standing commitment to robust cybersecurity infrastructure.