- Multi-turn jailbreak attacks succeeded nearly 93% of the time in testing eight major open-weight AI models.
- Enterprises may have a blind spot in assessing and deploying large language model safety.
Why It Matters for Texas Credit Unions
Texas CUs, like all financial institutions, need to ensure robust cybersecurity measures are in place to protect against potential AI-based threats.
Original Source Material
Cisco: One Prompt May Not Break Most AI Models, But a Conversation Will Cisco tested eight major open-weight artificial intelligence models and found multi-turn jailbreak attacks succeeded nearly 93% of the time, exposing a blind spot in how enterprises assess and deploy large language models safety.