Miscellaneous

Netflix has raised eyebrows by seemingly using AI-generated images in a documentary about Jennifer Pan’s murder-for-hire plot. The photos depicted Pan’s personality according to a friend, with telltale signs of AI manipulation evident. This use of generative AI to portray a real person in a true-crime show is concerning, especially when compared to instances of fictional AI-generated content in other media. Manipulating reality in this manner for entertainment crosses ethical boundaries and distorts the truth.

Roku is planning to show ads on its TVs even when you’re not using its apps, thanks to a new system that detects breaks in activity and overlays relevant ads. The technology would monitor HDMI signals for pauses, leveraging content recognition to display targeted ads based on what’s onscreen. While invasive, this tactic is common among big corporations, like Roku, known for their aggressive advertising strategies.

As AI companies grapple with a shortage of training data, the concept of “synthetic data” has emerged as a potential solution, although its effectiveness remains uncertain. Companies like Anthropic, Google, and OpenAI are striving to develop quality synthetic data, but challenges persist. Dubbed “Habsburg AI” by researcher Jathan Sadowski, this inbreeding of AI models could lead to unpredictable outcomes. With issues like “Model Autophagy Disorder” arising, the quest for reliable synthetic data continues, raising the question of whether AI systems can handle this unconventional approach.

Google is considering charging users for access to its experimental AI-integrated search feature, a potential move that would be a first for the tech giant. This shift could disrupt Google’s dominant ad revenue model, as AI-powered search might discourage users from clicking on ads, impacting the company’s baseline revenue. The proposal highlights the significant costs associated with AI systems and the challenges of balancing innovation with financial sustainability in the search marketplace.

Mayor Eric Adams defended the glitchy MyCity chatbot, acknowledging its flaws and the need for improvement, as it has been providing incorrect and possibly illegal information to users. The city officials are working on fixing the issues, with the chatbot now labeled as a beta product with potential inaccuracies. This incident joins a string of high-tech mishaps under Adams’ administration, raising questions about the efficacy and priorities of investing in such technology over more traditional and beneficial resources like supporting the city library system.