Generative AI: How Existing Regulation May Apply To AI-Generated Harmful Content

Among the many open questions about large-language models (LLMs) and generative artificial intelligence (AI) are the legal risks that may result from AI-generated content. While AI-specific regulation remains pending and continues to develop in jurisdictions around the world, the following article provides a high-level summary of illegal and harmful content risks under existing law, as well as mitigations that companies may wish to consider when developing baseline models and consumer-facing generative AI tools. (For copyright and intellectual property (IP)-related issues, seePerkins Coie Updates.)

Content Risks

Many content and communication platforms—such as social media platforms—engage in processes to find, block, and remove user content that is potentially harmful to others. This content moderation and related efforts to promote online safety are challenging and complex, particularly at scale. In recent years, lawmakers in the United States and around the