The Hidden Risks of Autonomous AI Agents | Dustin Allen x Shekhar Natarajan

In this episode of Tomorrow Today, Shekhar Natarajan and Dustin Allen explore AI safety, privacy and the future of autonomous agents.

1. Dustin Allen's journey and the story behind Trinitite
2. Why the rise of AI agents creates new safety challenges
3. Understanding agentic swarms and how multiple agents interact
4. Why probabilistic AI models can produce unpredictable outputs
5. The importance of context and human feedback in AI governance
6. How test driven governance helps identify AI edge cases
7. Deterministic inference and its role in controlling model behaviour
8. How smaller language models can improve reliability
9. Oracle guided distillation and model training
10. How Trinitite helps secure sensitive business workflows
11. The risks of AI agents handling employee and financial information
12. Why traditional restrictions may not prevent unwanted agent behaviour
13. Understanding personal boundaries in AI safety
14. Open source versus closed source AI models
15. Data governance and model transparency
16. The need for standards in agent to agent communication
17. How AI can affect personal privacy and trust
18. Why users should be careful about sharing sensitive information with AI
19. The growing adoption of personal AI agents
20. How Trinitite is making privacy controls more accessible
21. Why AI companies need better testing instead of simply collecting more data
22. The future of human autonomy in an AI driven world

Tune in to Tomorrow Today Show by SKN for powerful conversations with global changemakers, visionary leaders and extraordinary minds shaping our future.

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.