Panel Descriptions
The Practice of Law, ChatGPT, and Other Large Language Model (LLM) AI Advances
The first panel will examine the role of large language models (LLMs) technology like ChatGPT in legal practice. These advanced technologies very recently have acquired the capacity to perform various tasks that were once exclusively within the purview of lawyers, including drafting contracts, composing legal motions, preparing patent applications, and conducting legal analysis. Panelists will discuss the capabilities and limits of LLMs in law, both in the short term and nearer term. As well as the ethical issues surrounding the use of these technologies in law.
Generative Art, Music and Intellectual Property Law: Fair Use, Costless Creation, and Creativity
The second panel will delve into the issues surrounding the use of generative AI for art, text, and music creation. Until recently, AI systems could not produce creative output that were close to human output, but just in the past year AI advances now allow systems to create art, music, books, articles, and other creative output, at levels that meet, and sometimes exceed human output. This raises many novel issues in law, particularly in IP law, such as:
- Is it Fair Use under copyright law to take art, text, and other available information to train AI systems? Should it be?
- What about the implications for artists or other creators who may be displaced by this technology?
- How should IP Laws address the balance of rights in an era of costless creation?
- IP rights such as copyright were established during a time when producing high-quality human output, like art and music, required significant investment in time, resources, and talent. As a result, strong IP protections, like a 95-year copyright duration, were put in place to incentivize creativity. The cost of creating some high-quality, human-like creative works has significantly diminished due to AI, making it virtually costless to generate these outputs. Consequently, should IP protections for purely, or mostly, AI-generated outputs be unprotected or at least diminished?
Potential Positive and Negative Impacts of Large Language Models on Governance and Society
The final panel will explore the broader societal considerations of LLM technology and its potential impact on governance and society from both a positive and negative perspective. Topics of focus will include:
- How AI-generated legal documents may increase access to legal services to the under-served, while also maintaining quality and ethical standards.
- How AI will likely increase overall productivity for society, but also displace many current workers, and how to ensure that these productivity gains are equally distributed among society, particularly to those whose jobs are displaced by the technology.
- How AI will be disruptive in terms of many areas, including security, privacy, and government operations.
- Generative AI holds the potential to create near-costless, high-quality text that could significantly impact governmental processes.
- For instance, in the past, drafting high-quality legal text, such as patent applications or comments for “Notice and Comment” in administrative rule-making, required the expertise of a lawyer and consumed considerable time and resources. However, with the advent of AI, the creation of reasonable quality legal text (which is fairly good, although not quite at the level of an experienced lawyer), but at least plausibly so, has become almost effortless and cost-free. This development raises concerns about the possibility of governments being inundated with seemingly high-quality patent applications, notices-and comments, etc., which has been produced by AI at virtually no cost. In the past, government officials could recognize, and easily dismiss, low-quality texts.
About Generative AI
General Explainers for Generative AI:
- NY Times Article — How ChatGPT Kicked Off an A.I. Arms Race
- Stephen Wolfram Writings — What Is ChatGPT Doing…and Why Does It Work?
- McKinsey & Company Article — What is generative AI?
Professor Harry Surden, Colorado Law Professor and Silicon Flatirons AI Initiative Director, demonstrates ChatGPT:
- ChatGPT Generates a Federal Motion to Dismiss using GPT-4
- ChatGPT Applies Copyright Law