An attacker can inject indirect prompts to trick the model into harvesting user data and sending it to the attacker’s account.
The post Claude AI APIs Can Be Abused for Data Exfiltration appeared first on SecurityWeek.
New AI Safety Initiative Aims to Set Responsible Standards for Artificial Intelligence
Major software vendors sign on to a new security initiative to create trusted best practices for artificial intelligence deployments.
The post New AI Safety Initiative Aims to Set Responsible Standards for Artificial Intelligence appeared first on SecurityWeek.