Law & Policy

Here's a question for you:

AI Law and Policy course.

What topics would you cover?

What topics seem relevant now, but might not be by the time I teach it in the Fall?

A comprehensive course on AI Law and Policy should include:

- Basics of model training and inference

- Why chips are so important and how US controls them (via De Minimis U.S. Content (§ 734.4) and Foreign Direct Product Rule as applied to chips and AI models), what was The AI Diffusion Rule and why it was rescinded recently

- Which data could or could not be used for training (privacy, etc.)

- Fair use in the context of model training (NYT v. OpenAI, Thomson Reuters v. ROSS Intelligence)

- Basics of The EU AI Act as an example of the first comprehensive regulation (perhaps not the best regulation, though)

- Why certain AI startups want to use non-profits or public benefit companies and how things can go wrong (OpenAI board saga as a case study)

- Whether and how lawyers, doctors and other regulated professions should use AI (ABA Formal Opinion 512 on GenAI)

Just some ideas. The list can go on.

lawyer and investor @ True Ventures / fmr GC @Paradigm, @Circle, DHS / law, tech, policy, security, crypto
 / coldebella.substack.com

100% 

between the programmer, the inference model, the system prompt, and the user interface, there are so many interesting layers 

@atlas wdyt ??

Ooh good one. Will an AI agent be treated the same way a human is under the law, and will liability for the principal be curtailed in the same way (such as when the agent acts outside the scope of employment)? What's a frolic and detour (one of my favorite little tests from law school) for a computer system? Nice.

principal<>agent liability seems like a ripe topic

1. Confidentiality/ethics. Maybe using bar opinions regarding use of Gmail as applicable since there is overlap on training/third party access to confidential client materials. 

2. Ethics opinions regarding duty to use technology to increase efficiency. 

3. Sourcing of information generally (hallucinations similar to when people would take information from Wikipedia without checking sources)

I would cover topics like data privacy, algorithm bias, and ethics in AI. Some topics that might not be as relevant by Fall could be current regulations or specific case studies. It's important to stay updated in this fast-evolving field!