The Current Legal Climate Surrounding AI and what You Need to Do About It
Today’s AI systems are using Large Language Models (LLMs) to generate content. They are also being...
Today’s AI systems are using Large Language Models (LLMs) to generate content. They are also being sued for those LLM’s overreach. The three stories below have all taken place since our last newsletter:
• Two US. authors, Paul Tremblay and Mona Awad, have filed a lawsuit against OpenAI, alleging that their works were used without permission to train the AI system, ChatGPT. The lawsuit claims that ChatGPT mined data from thousands of books, infringing on the authors’ copyrights. It is estimated that OpenAI’s training data incorporated over 300,000 books, some from illegal sources offering copyrighted books without permission. The authors argue that ChatGPT can generate accurate summaries of their books, indicating their presence in its database. The lawsuit seeks unspecified damages for a nationwide class of copyright owners whose works OpenAI allegedly misused.
• UK Prime Minister Rishi Sunak revealed that the government is considering regulations to mandate “watermarks” on AI-generated content to prevent public deception. US President Joe Biden has also endorsed the idea of compulsory watermarks. The UK aims to lead global AI regulation, balancing the EU’s restrictive approach and the US’s laissez-faire attitude.
• Three authors; Sarah Silverman, Richard Kadrey, and Christopher Golden have joined into a class-action lawsuit with over 1,000 others against OpenAI and Meta (Facebook) alleging that their copyrighted material can be used to create derivative works and “cannot function without the expressive information extracted from the Plaintiff’s Infringed Works”.
We “compliance folk” need to come up to speed. Here are some questions you need to ask yourself:
- What are the current licensing standards and international laws regarding machine readability?
- Is it possible to obtain permission for machine readability through licensing?
- What are the rights to the tagged data and trained models once completed?
Here’s what to do about it:
1. Read our whitepaper to answer those three questions.
2. Make a license. We have worked directly with Perkins Couie to create a Generative Commons License that you can fill out and download for your business. Fill it out and route it around so that you are protected!
3. Share this video. Share a fun 1.5-minute video for education about this topic.