A Tough Week for Social Media, Sora Says Goodbye, and AI Cautionary Tales

Two court rulings against Meta (parent company of Facebook and Instagram) landed this week including one where Mark Zuckerberg (Meta CEO) and Adam Mosseri (Head of Instagram) previously testified. The New Mexico verdict included a penalty of $375 million. New Mexico Attorney General Raúl Torrez released a statement saying, “Meta executives knew their products harmed children, disregarded warnings from their own employees, and lied to the public about what they knew.”

• 🎙️ When following news regarding social media regulations, safety features or accountability we always ask, “Yeah, but are they changing their business model?” At this point, neither verdict mandates such changes, but the judge in the New Mexico case will make a “public nuisance” ruling in May which might require Meta to implement changes or programs to address harms.

OpenAI announced it was shutting down Sora, its AI video generator app and Disney abandoned a $1 billion licensing deal with the tech company. With dwindling app downloads, compute costs, and financial struggles, OpenAI is prioritizing endeavors that raise revenue (like ads).

• 🎙️ Our two cents? In a world that’s increasing artificial, people long for truth and authenticity.

Yesterday the Science journal posted a study comparing the responses of AI chatbots to human replies to Reddit posts. It states, “The model’s responses were nearly 50% more sycophantic than humans, even when users engaged in unethical, illegal, or harmful behaviors.”

• 🎙️ Two of Know Curtains core principles come into play: Transparency and Integrity. It’s important to not only know AI’s limitations, but also to be honest about how we engage with technology. The study notes that “users preferred and trusted sycophantic AI responses.” Human nature seeks answers and advice that we want to hear, not what we need to hear. Find AI models or modify prompts that allow you to challenge your own perspective.

Two teenage boys were sentenced to probation and 60 hours of community service by a Pennsylvania judge for using AI to create explicit images of girls in their school. Judge Leonard Brown told the boys they would likely be in prison if they were adults.

• 🎙️ State laws on non-consensual intimate images (NCII) vary. In this case, the victims were minors, with the boys admitting to 59 felony counts of sexual abuse of children. The TAKE IT DOWN Act, signed into law last year makes knowingly sharing or threatening to share NCII a federal crime. Bottom line? Don’t generate or share authentic or AI-generated nudes.

Driver’s Training for the Online World has entire module focusing on AI. If anyone wants to try before they buy, test drive the entire first module for free.

• 🎙️ How does it work? Each module has five lessons. It starts with a 10-12 minute video, and it’s followed by a quick quiz, discussion questions, thought experiments (real-life what-would-you-do scenarios), and action ideas. It’s great for individuals and the perfect on-ramp for family discussion and planning. Use the code “GREENLIGHT” to get 25% off.