Talk
GenAI - Chances and Risks | DSC Adria '24
Explore Generative AI's potential and challenges, from enterprise use to risks like prompt injections, with insights on best practices of AI governance.
Kristian Kamber
DATE
Jul 5, 2024
TIME & LENGTH
25 min
STATUS
Available on demand
LANGUAGE
English
This talk at DSC Adria '24 of Kristian Kamber, Co-Founder and CEO of SplxAI, explores the transformative potential of generative AI, from automating tasks to advanced communication and decision-making. Key topics include its current enterprise applications, such as enhancing customer service and automating administrative tasks, alongside security challenges like prompt injections, data leaks, and production readiness. Kris shares insights on securing AI systems through penetration testing, input/output validation, and collaborative team efforts. The talk highlights the rapid adoption of GenAI in emerging markets, the need for continuous security updates, and the pressing demand for responsible AI governance to unlock its full potential.
Securing Generative AI: Strategies to Combat Prompt Injection and Data Leaks
Brand Reputation Risks: AI vulnerabilities can lead to incidents like chatbots recommending competitors, undermining trust and damaging brand image.
Proactive Security Measures: Penetration testing, input/output validation, and third-party guardrails are vital for protecting AI systems in production.
Emerging Threats: Invisible characters, multimodal attacks, and governance gaps highlight the ongoing need for vigilance and continuous improvement.
