Generative AI, which uses large language models (LLMs) such as GPT, can have various vulnerabilities that need attention.

Some common concerns include adversarial attacks, wherein attackers manipulate input data to cause the model to produce incorrect or harmful outputs; and data privacy, wherein LLMs may unintentionally memorize and output sensitive information from their training data, posing a risk to user privacy and breach of legal compliances such as GDPR; and additional security concerns related to the leakage of sensitive data. Furthermore, issues related to safety, bias, misinformation, hallucinations, lack of transparency and explainability need to be addressed.

The session will delve into the crucial role of CISOs in safeguarding organizations, including AI regulations and compliance requirements, establishing AI ethical guidelines, employee training, and conducting risk assessments to identify potential threats and vulnerabilities in AI systems, and putting in place security controls and guardrails.