Topic: Safe Harbor

XAI filed a lawsuit to enjoin a new California law requiring AI companies to release summaries of their data training sources. The legislation aims to address the "AI porn problem" and misinformation by forcing transparency in model development. Experts suggest a "safe harbor" provision is needed to allow researchers to "red team" models—acting as malicious users to find exploits—without fear of legal prosecution.

The Trump campaign announced plans to prosecute cases in federal court starting Monday, focusing on the distinction between "legal ballots" and "fraudulent" ones. Legal experts discuss the "Safe Harbor" deadline of December 8th and the plenary rights of state legislatures to award electors if results remain disputed. The hosts anticipate the involvement of Justice Amy Coney Barrett as an "originalist" on the Supreme Court.

The discussion turns to the legal risks faced by social media companies like Twitter and Facebook as they begin to act as editors rather than common carriers. By editorializing content, they may lose "safe harbor" protections under the DMCA. The hosts also mock Uber's refusal to obtain autonomous vehicle permits in California and note the AP Stylebook's new requirement to call these companies "ride-hailing" rather than "ride-sharing" services.

The European Court of Justice strikes down the Safe Harbor Agreement, giving U.S. tech companies three months to negotiate a new deal for transferring data from the EU. This ruling specifically impacts Facebook's ability to analyze European user data in the United States. Simultaneously, the Volkswagen emissions scandal deepens with reports that the company used multiple versions of "defeat devices" over a prolonged period to fool regulators.