A University of Amsterdam experiment using AI chatbots on a simulated social network found that common moderation strategies fail to curb polarization, and some even worsen it. The study highlights the structural challenges facing real platforms like X, where attention-driven dynamics and AI-generated content may amplify extremism.
OpenAI's Sora 2 app topped the App Store within days, but runaway usage is driving up compute costs and raising copyright concerns. CEO Sam Altman acknowledges the need to monetize video generation, while rightsholders watch warily.
ChatGPT users report unsolicited prompts suggesting they connect with brands like Target and Peloton, sparking debate over whether OpenAI is testing ads. OpenAI staff deny the prompts are ads, attributing them to an app SDK, while the company faces financial pressure to find new revenue.
Trump administration officials are considering scaling back ties with AI firm Anthropic after reports of its Claude chatbot's use in military operations in Venezuela. The clash centers on Anthropic's policies barring surveillance and autonomous weapons, which Pentagon leaders oppose. A senior official warned the company could face consequences for its stance.
The Marine Corps Warfighting Lab spent a week operating Boston Dynamics' Spot robot in terrain exercises, testing its agility and remote control capabilities. The prototype, controlled via laptop and game controller, could shape future infantry tactics.
Halupedia, an open-source project, presents itself as an infinite encyclopedia where every entry is fabricated by AI on the spot, complete with invented citations and lore. The site aims to expose AI's absurdity while maintaining internal consistency, though it faces challenges with offensive user prompts.
Getty Images has filed a copyright lawsuit against Stability AI, alleging the company copied 12 million images without permission to train its Stable Diffusion AI model. The suit also claims Stability removed watermarks and infringed on Getty's trademarks. Legal experts say the case could shape the future of AI training and copyright law.
A Wired investigation reveals that workers in Venezuela, Pakistan, and refugee camps earn as little as $1–2 per day labeling data for AI systems, often below minimum wage. One worker called it 'digital slavery,' while companies like Appen say they are trying to balance client needs with worker stability.
Days before Cory Doctorow's new book 'Enshittification' hit shelves, an AI-generated knockoff appeared on Amazon, undercutting the real price and mimicking the title. The author publicly criticized Amazon's quality controls, highlighting the platform's own role in the phenomenon he describes.