On June 14th, a Senate bill was introduced by U.S. Senators Josh Hawley and Richard Blumenthal, representing Missouri and Connecticut respectively. The bill seeks to revoke the special privileges currently granted to artificial intelligence (AI) companies, which are presently similar to the protections enjoyed by online computer services providers under the Communications Decency Act of 1996 (CDA).
The bipartisan bill aims to establish accountability for companies in addressing harm; however, it remains uncertain whether Section 230 is applicable to artificial intelligence (AI).Section 230, located in Title 47, Section 230 of the Communications Decency Act (CDA), grants online service providers protection against liability for user-generated content. It also offers immunity from legal action for hosting illegal content, as long as sincere efforts are made to promptly remove such content upon identification.
Critics of Section 230 claim that it relieves social media platforms and online service providers from the obligation of being responsible for the content they host. The U.S. Supreme Court recently made a decision not to modify Section 230 in response to a lawsuit in which plaintiffs attempted to hold social media companies liable for damages resulting from the alleged hosting and promotion of terrorist-related content on their platforms.
According to the court’s ruling, a social media site cannot be held responsible for the recommendations generated by its algorithms, just as email or cellular service providers cannot be held responsible for the content transmitted through their services.
During a recent Senate hearing, Sam Altman, the CEO of OpenAI, expressed his belief to U.S. Senator Lindsey Graham that Section 230 did not apply to his company. When questioned by Senator Hawley about a hypothetical scenario in which Congress enabled individuals harmed by AI to testify in court, Altman responded, “Please excuse my lack of knowledge, but can’t people sue us?”. Although Section 230 does not explicitly address generative AI, the discussions regarding its applicability to generative AI technologies may hinge on the interpretation of the term “online service.”