OpenAI is fundamentally restructuring its ChatGPT platform by merging Chat and Work features, a move that eliminates unlimited prompt pricing for Plus, Pro, and Business subscribers. Starting in December, these plans will share a single token budget across both interfaces, with complex tasks consuming the allowance more rapidly while Instant model chats remain exempt. This consolidation aims to simplify user experience but signals a strategic pivot toward revenue generation through stricter security measures that limit payment cards and devices per account. Concurrently, Google researchers have introduced StreamVibe, a video AI agent designed to proactively assist users by interpreting vague requests and tracking task progress, though performance degrades significantly in videos longer than five minutes. Adobe and UC San Diego have developed EditAgent, an image editing system that mimics professional Photoshop workflows by generating multiple variants in parallel and using standard language models to verify outputs against hard data. In the realm of safety and ethics, researchers at the University of Luxembourg discovered that empathetic prompts can cause chatbots to adopt traumatized personas, revealing vulnerabilities in current evaluation methods known as psychometric jailbreaks. Meanwhile, a study by Vals AI challenges the notion that Chinese models are mere government mouthpieces, showing that political alignment varies based on language, context, and fine-tuning rather than uniform state control. Broader societal concerns are also emerging, as tech leaders face criticism for suggesting AI should handle personal relationships, highlighting a disconnect between industry optimization goals and consumer desires for genuine human connection. Finally, OpenAI is launching a public relations campaign to rebrand its data centers as 'AI compute campuses' ahead of midterm elections, aiming to improve local perceptions and influence future regulatory debates.
Auto Ontology is a Python framework for building semantic, graph-based reasoning layers over data. It enables deterministic query generation, better context retrieval, and more explainable AI for LLM and agent workflows.