Safe Mode Podcast What will GenAI regulation look like in 2025?
Nov 7, 2024
In this conversation, Casey Bleeker, CEO of SurePath AI, a leader in generative AI security, delves into the evolving landscape of AI regulation in the U.S. He discusses the potential for varying regulations across states, akin to current data privacy laws. The challenges of creating a cohesive national strategy are highlighted, along with the need for robust governance frameworks for generative AI. Further, Casey emphasizes the importance of transparency and consumer protection, addressing how companies can balance innovation with compliance.
AI Snips
Chapters
Transcript
Episode notes
Unintentional Data Leakage
- A customer trained their entire data warehouse into Llama 70 billion without realizing the implications.
- This gave all users unfettered access to sensitive internal enterprise data, highlighting data leakage risks.
The Double-Edged Sword of AI in Cybersecurity
- Blocking AI is ineffective; employees find workarounds or use personal devices, increasing risk.
- Generative AI increases the threat kill chain, enabling autonomous zero-day attacks and sophisticated social engineering.
Importance of Audit Trails
- Maintain an audit trail of AI interactions to establish an affirmative defense in legal or regulatory situations.
- This proactive approach minimizes effort in mounting a defense and protects against civil suits.