AI & Freedom of Speech

In their 2016 paper, Siri-Ously? Free Speech Rights and Artificial Intelligence, and its 2017 follow-up, SIRI-OUSLY 2.0: What Artificial Intelligence Reveals About the First Amendment, Massaro & Norton argue that speech generated by artificial intelligence (AI) is protected under the First Amendment, based on free speech theory and doctrine. Their position is anchored in the evolving focus of First Amendment jurisprudence on the value of listeners and the desire to limit government suppression of speech. Although their work provides a thorough analysis of AI's challenges, it did not fully anticipate the emergence of AI using large language models (LLM) with natural language understanding (NLU), communicating in human-like interactions. These AI-generated communications are often indistinguishable from human speech and are governed by private corporate protocols. Consequently, the boundaries between human and AI speech have become increasingly blurred, raising questions about authorship, ownership, and responsibility. This paper argues that while AI is not yet an independent rights-bearer, the proliferation of public access to AI, and the intertwinement of human and augmented speech, necessitates a reexamination of ownership under free speech theory and doctrine. Exploring the relationship between the creator of the original work and AI's involvement in the creation process, I aim to provide a clearer understanding of the implications of AI-generated speech in the contemporary digital landscape.