BNP Paribas SA is working with French AI startup Mistral AI and other partners to prepare for cybersecurity risks associated with new AI model capabilities, according to reports referencing Bloomberg. The collaboration is described as a preparatory effort aimed at addressing threats related to models such as Anthropic PBC’s “Mythos.”
The coverage characterizes the project as part of a broader push among European institutions to develop or access AI cybersecurity tools and capabilities that are already used by some US peers. It also notes that European supervisors may face limitations in obtaining these tools reliably, creating a motivation for regional development.
While the reports emphasize that the discussion around “Mythos” includes whether such models are accessible or not, they also stress the need to consider downstream security implications regardless of access. The initiative focuses on readiness for potential misuse or risk scenarios stemming from advanced AI models, bringing together a major European bank, a local AI company, and additional partners.
