May 21, 2025
Link to TJLP Press Release (reproduced below).
We applaud Judge Conway for her thoughtful and nuanced opinion today, allowing our client’s claims to go forward against defendants Character.AI, its co-founders Noam Shazeer and Daniel DeFreitas, and Google.
In the decision, Judge Conway cut through the defendants’ faulty legal reasoning to reach the core issue in this case: whether the large language model (LLM) that powers Character.AI’s social chatbots is a defective product, one that results in foreseeable harms to young people like Sewell that all four defendants had a legal duty to prevent.
This decision is truly historic. It sends a clear signal to companies developing and deploying LLM-powered products at scale that they cannot evade legal consequences for the real-world harm their products cause, regardless of the technology’s novelty. Crucially, the defendants failed to convince the Court that those harms were a result of constitutionally-protected speech, which will make it harder for companies to argue so in the future, even when their products involve machine-mediated “conversations” with users.
The decision has several important takeaways:
- Character.AI’s LLM is a product subject to product liability law and not merely a service that would not be.
- The First Amendment does not apply here, giving LLM-powered chatbots a get-out-of-jail-free card.
- While Character.AI has the ability to assert the First Amendment rights of its users on their behalf, the outputs of the LLM were nothing more than words strung together by probabilistic determinations. The outputs are not protected speech, as they lack the human intention required for expression.
- The judge relied on Justice Barrett’s skepticism in her Moody concurrence over whether algorithmic outputs are protected under the First Amendment when there is no human intervention.
- Beyond Character.AI, Google and the individual co-founders must also remain defendants in the case.
- We sufficiently alleged that Google gave significant assistance to launching and sustaining Character.AI’s operations and also had actual knowledge of the harms inherent to the social chatbot product.
- Shazeer and DeFreitas, the co-founders, were subject to the Court’s jurisdiction due to their personal involvement in the product at issue.
- The court’s refusal to dismiss Shazeer and Defreitas from the case is especially notable, with implications for future cases in which company founders are instrumental to the harms caused by a tech product.
In a climate where generative AI companies are being incentivized to move fast and encouraged to break things in the name of innovation, we sincerely hope that this decision will force Silicon Valley to pump the brakes. Judge Conway’s ruling suggests that proceeding with “business as usual” will no longer be tolerated — neither for tech companies nor their founders and developers.