At Norm Ai, we are implementing a deep understanding of law, the language of alignment, into specialized Legal & Regulatory AI systems. The integration of law and AI is becoming increasingly important as AI capabilities advance, and AI is more widely deployed. While there have been suggestions to integrate ethics with AI to increase alignment with humans, it is unclear how to determine these ethics and who gets a say in the process. Instead, we propose that the target of AI alignment should be democratically endorsed law, which provides a legitimate grounding for AI behavior and can serve as a set of methodologies for conveying and interpreting directives and a knowledge base of societal values.

No one can directly specify “good” AI behavior up front. It’s far too complicated to do for all relevant future scenarios. Similarly, parties to a legal contract cannot foresee every contingency, and legislators cannot predict all the specific circumstances under which their laws could be applied. Law, as the applied philosophy of multi-agent alignment, can uniquely serve as an AI goal specification framework. Methodologies for making and interpreting law, which advance shared goals in new circumstances, have been refined over centuries.

One of our important goals at Norm Ai is to build the specialized Legal and Regulatory AI agents for AI agent guidance to follow the spirit of the law. John Nay, Norm Ai founder and CEO, lays out his vision and our plan to implement it, in an article published with Stanford Law: “Aligning AI Agents with Humans through Law as Information”

Read the full article at the link below:

https://law.stanford.edu/publications/aligning-ai-agents-with-humans-tehrough-law-as-information/.