Manage risk in AI deployments: Deep-dive into OWASP Top 10 for LLM
Watch On Demand
Summary
Organizations are deploying AI rapidly for business needs, but often the security aspects of the deployments are overlooked. Organizations need to consider aspects around sensitive information disclosure, supply chain risks, data and model poisoning. Attackers would look to target the sensitive data used for training, the vulnerabilities in the applications and models, or carry out attacks that could manipulate the behavior of the models.
Thus, it is important to leverage assessment frameworks to benchmark your AI deployment. One such common framework is OWASP Top 10 for LMM. OWASP Top 10 for LMM lists the top 10 most critical vulnerabilities often seen in LLM applications selected by a large community of security leaders, highlighting their potential impact, ease of exploitation, and prevalence in real-world applications. These vulnerabilities range from prompt injection to supply chain vulnerabilities to unbound consumption.
- Listen this session to hear experts talk about:
- Common risks associated with AI deployments
- Overview of the OWASP Top 10 for LLM framework
- Top 10 vulnerabilities, as identified by the framework
- How Guardium AI Security can help, and a demo.
Key speakers
Assaf Miron
Senior Product Lead, Cloud & Emerging Capabilities, IBM Guardium
Anshul Garg
Senior Product Marketing Manager, IBM Guardium
Manage risk in AI deployments: Deep-dive into OWASP Top 10 for LLM
Watch On Demand
Summary
Organizations are deploying AI rapidly for business needs, but often the security aspects of the deployments are overlooked. Organizations need to consider aspects around sensitive information disclosure, supply chain risks, data and model poisoning. Attackers would look to target the sensitive data used for training, the vulnerabilities in the applications and models, or carry out attacks that could manipulate the behavior of the models.
Thus, it is important to leverage assessment frameworks to benchmark your AI deployment. One such common framework is OWASP Top 10 for LMM. OWASP Top 10 for LMM lists the top 10 most critical vulnerabilities often seen in LLM applications selected by a large community of security leaders, highlighting their potential impact, ease of exploitation, and prevalence in real-world applications. These vulnerabilities range from prompt injection to supply chain vulnerabilities to unbound consumption.
- Listen this session to hear experts talk about:
- Common risks associated with AI deployments
- Overview of the OWASP Top 10 for LLM framework
- Top 10 vulnerabilities, as identified by the framework
- How Guardium AI Security can help, and a demo.
Key speakers
Assaf Miron
Senior Product Lead, Cloud & Emerging Capabilities, IBM Guardium
Anshul Garg
Senior Product Marketing Manager, IBM Guardium