Ethical Machines How Do You Control Unpredictable AI?
5 snips
Jul 10, 2025 
Walter Haydock, a former national security policy advisor and founder of StackAware, dives into the intriguing complexities of unpredictable AI. He discusses the dual nature of large language models—capable of both creativity and chaos. Haydock emphasizes the critical need for structured risk assessments to navigate the pitfalls of integrating agentic AI into organizations. He highlights dangers like data poisoning and calls for stricter testing and monitoring to ensure responsible AI deployment while fostering innovation.
AI Snips
Chapters
Books
Transcript
Episode notes
Unpredictability of Agentic AI
- Agentic AI systems operate unpredictably, complicating testing and control.
- This unpredictability underpins their creativity but poses significant cybersecurity risks.
Agentic AI Database Risks
- Walter gave an example where agentic AI hooked to a database might misinterpret commands.
- It could accidentally modify or delete unintended data, causing serious issues without proper controls.
Unpredictability as Core Challenge
- The core challenge is agentic AI's unpredictable behavior, not just agency itself.
- This unpredictability resembles human error and requires similar risk thinking.
