Legal News for Weds 8/28 - Girardi Guilty, EU-US Split on AI Privacy, Trump Indictment Updated by Smith, TikTok Lawsuit Over Blackout Challenge

Minimum Competence

This Day in Legal History: Alabama Ten Commandments Monument

On August 28, 2003, the Supreme Court of Alabama took down a monument of the Ten Commandments from its courthouse rotunda, marking the culmination of a high-profile legal battle. The monument had been installed by Chief Justice Roy Moore in 2001, who argued that it reflected the moral foundation of U.S. law. However, this act sparked a federal lawsuit, Glassroth v. Moore, in which three Alabama attorneys claimed the monument violated the Establishment Clause of the First Amendment, which prohibits government endorsement of religion.

The federal District Court for the Middle District of Alabama agreed with the plaintiffs, ordering Moore to remove the monument. Moore refused, maintaining that he had a duty to acknowledge God in his official capacity. The case was subsequently appealed to the Eleventh Circuit, which upheld the lower court's ruling. When Moore continued to defy the court orders, the Supreme Court of Alabama intervened, removing him from his position as Chief Justice. This case became a significant moment in the ongoing debate over the separation of church and state in the United States.

It is worth noting that Roy Moore, the then-Chief Justice of the Alabama Supreme Court who so vociferously argued for the inclusion of the Ten Commandment monument is the selfsame Roy Moore that, during his 2017 U.S. Senate campaign, saw nine women accuse him of inappropriate conduct. Three of the women claimed they were assaulted by Moore when they were aged 14, 16, and 28. The other six women described Moore pursuing relationships with them when they were as young as 16. Independent witnesses corroborated that Moore had a reputation for approaching teenage girls at a local mall. Moore's responses to the allegations were inconsistent, initially recognizing some accusers but later denying knowledge of any of them.

Thomas V. Girardi, a prominent figure in toxic tort litigation, was convicted on four counts of wire fraud in Los Angeles federal court. Once renowned for his work on the Erin Brockovich case and his appearances on "Real Housewives of Beverly Hills," the disbarred attorney faced accusations of defrauding vulnerable clients. The jury reached a unanimous verdict after just four hours of deliberation, rejecting Girardi's defense that his cognitive decline prevented him from forming intent to commit fraud.

Prosecutors argued that Girardi knowingly deceived clients, fabricating excuses to explain the missing funds, which he had already spent. The trial centered on the suffering of clients who were betrayed by Girardi in their darkest moments, leading to their financial and emotional devastation. Girardi could face up to 80 years in prison at his sentencing in December. His former CFO, Christopher Kamon, will also stand trial for related charges. The case highlights Girardi's history of evading disciplinary action despite numerous complaints and reveals potential future charges against other senior lawyers at his firm.

Thomas Girardi Found Guilty by Jury of Defrauding Clients (2)

A recent decision by a German privacy regulator has sparked intense debate about how personal data is handled by AI models like large language models (LLMs). The Hamburg Commissioner for Data Protection concluded that LLMs, despite generating personal data, do not store such information in a way that makes it identifiable, challenging the notion that AI systems can retain personal data.

This stance contradicts findings by technologists who argue that LLMs can memorize and reproduce specific data, including personal details. The German position could limit individuals' ability to control their data in AI systems, potentially leading to significant differences in how the U.S. and the EU regulate AI. While California is pushing for laws that explicitly protect personal data in AI, the German approach may set a precedent for a more lenient interpretation under the GDPR. This divergence highlights the complexity of applying traditional privacy laws to AI technologies, with ongoing discussions about how to reconcile these differing perspectives.

By way of brief background, LLMs do not directly memorize the training material they are exposed to. Instead, they analyze vast amounts of text data and learn patterns, correlations, and structures within the language, which are then used to generate responses. This learning process involves creating a complex mathematical representation of language—a model—rather than storing specific pieces of text verbatim. However, because these models are trained on enormous datasets, they might sometimes generate outputs that resemble specific phrases or data points encountered during training, especially if those phrases are common or particularly distinctive. This can occasionally lead to the unintentional reproduction of personal or sensitive information from the training data, even though the model itself does not store or recall such information in a traditional, deliberate sense.

Of course, that would all be of slim comfort to someone who sees an AI chatbot spit out their home address and social security number in response to a prompt.

Personal Info in AI Models Threatens Split in US, EU Approach

Special Counsel Jack Smith is moving forward with prosecuting Donald Trump for allegedly attempting to overturn the 2020 election, despite a recent setback from the Supreme Court. The court found that Trump might have partial immunity from prosecution for actions taken as president, leading Smith to file a revised indictment. This new version removes claims related to Trump’s communications with government officials, including efforts to involve the Justice Department, but retains the core charges accusing Trump of conspiring to reverse his election loss.

The case comes as Trump campaigns for the 2024 election, adding tension to the legal proceedings. Trump criticized the indictment on social media, calling for its dismissal. The updated indictment also cuts references to former Justice Department official Jeffrey Clark as a co-conspirator and modifies Trump’s description, downplaying his role as president at the time. The case now focuses more on Trump’s role as a candidate rather than his presidential actions. As the case progresses, Trump faces other legal challenges, including cases involving classified documents and charges in Georgia related to the 2020 election.

Trump Special Counsel Presses Ahead With 2020 Election Case (3)

A U.S. appeals court has revived a lawsuit against TikTok by the mother of a 10-year-old girl who died after attempting a dangerous "blackout challenge" promoted on the platform. The Philadelphia-based 3rd U.S. Circuit Court of Appeals ruled that TikTok is not shielded by Section 230 of the Communications Decency Act, which typically protects internet companies from liability for user-generated content.

The court found that Section 230 does not apply when TikTok’s algorithm actively recommends harmful content, viewing such recommendations as the company’s own speech. This decision marks a shift from previous interpretations of Section 230, which had generally protected platforms from liability for failing to prevent the spread of harmful content. The ruling overturns a lower court's dismissal of the case, allowing the mother, Tawainna Anderson, to pursue claims against TikTok and its parent company, ByteDance, following her daughter Nylah's death in 2021. The case could have significant implications for how tech companies are held accountable for the content their algorithms promote.

TikTok must face lawsuit over 10-year-old girl's death, US court rules | Reuters

This is a public episode. If you’d like to discuss this with other subscribers or get access to bonus episodes, visit www.minimumcomp.com/subscribe