The author proposes three approaches to accelerate the industry's time to market with these new technologies: schema consistency across products, risk as an outcome measure, and user experience as the key to unlocking user interaction data. The author emphasizes the importance of data cleanup, the use of risk measurement as a reward function in RLHF models, and the need for UX teams to capture user activity in the same way e-commerce sites do.
Key takeaways:
- The future of AI/ML in security lies in the combination of large language models (LLMs) and reinforcement learning from human feedback (RLHF), which can improve AI-generated text and align it with human preferences.
- Security is a language that needs to be encoded into LLMs, requiring the training of new neural networks that can perform specific security analyses or tasks.
- Reinforcement learning through human feedback is crucial, with the opportunity to systematically gather data about end users’ behavior and use that data as the human in the loop.
- Three approaches to accelerate the industry's time to market with these new technologies include schema consistency across products, risk as an outcome measure, and UX as the key to unlocking user interaction data.