Goldman Sachs is taking a cautious, proactive stance toward Anthropic’s advanced AI model Mythos because of its ability to autonomously discover and exploit software vulnerabilities that could significantly disrupt financial systems. The bank is collaborating with Anthropic, cybersecurity partners, and other major firms through Project Glasswing to assess risks and strengthen defenses before wider deployment. #Mythos #Anthropic
Keypoints
- Goldman Sachs is closely monitoring Mythos and other next-generation AI models for cybersecurity implications.
- Mythos can autonomously find and exploit sophisticated software vulnerabilities, sometimes outperforming human experts.
- Anthropic has restricted access to Mythos and shared it selectively via Project Glasswing with major firms including Goldman Sachs.
- Banks are especially exposed due to complex mixes of modern and legacy systems that AI could target.
- Financial institutions are engaging early with model publishers and investing in defenses to balance risk and potential operational benefits.
Read More: https://thecyberexpress.com/goldman-sachs-ai-risks-anthropic-and-mythos/