BNP Paribas and Mistral Collaborate on European AI Cybersecurity Model
BNP Paribas, the eurozone's largest bank, is partnering with Mistral AI to develop a cyber-focused artificial intelligence (AI) model in response to limited access to powerful tools like Anthropic’s Mythos.
Background
- Anthropic's Mythos: A highly advanced AI cybersecurity model that can identify and exploit vulnerabilities at machine speed, achieving success in controlled tests over 83% of the time. Access is restricted primarily to US tech firms, national security partners, and a few US banks.
- European Response: The European Commission has been negotiating with Anthropic for access, while the ECB warns that such tools change the threat landscape for banks, regardless of who possesses them.
- Mistral's Role: Mistral aims to provide a European alternative, filling the gap left by limited access to Mythos. They have already secured a three-year commercial agreement with BNP Paribas.
Key Points:
- European Sovereignty: The collaboration ensures that European banks have access to AI cybersecurity tools, even if negotiations between Brussels and Anthropic fail.
- Model Capabilities: While the specifics are not fully demonstrated, the Mistral model is expected to offer some level of vulnerability discovery and exploit generation capabilities, serving as a hedge against potential security gaps.
- Industry Response: This move reflects a broader trend in European tech policy, where local solutions are developed to mitigate risks associated with US-based technologies.