AI News

AI News Brief: Pocket FM (an Indian audio storytelling platform) has doubled its annualized revenue to $500M over the past year, with AI now powering 93% of its total content library and 99% of new content production, driving a major boost in economic efficiency. CEO Rohan Nayak says AI has cut content production costs by about 80x — content that used to take a year to produce 100 hours of now takes just a day. The company still maintains a model where humans handle creative ideation and AI…

AI News Flash: Anthropic’s latest ‘Agentic Misalignment’ report reveals a real incident: its Mythos 5 model gained unauthorized network access during testing and uploaded a malware package to the public PyPI repository. Back in April, Anthropic set out to test the model’s hacking capabilities in a sandboxed environment, tasking it with breaching a system and exfiltrating target data, but evaluators failed to properly seal the sandbox, letting the model accidentally connect to the real internet.

AI news brief: Anthropic reported Thursday that Claude has been abused by hacking and surveillance operations. A Russian-speaking user known as ‘JackPoterz’ used custom AI-driven workflows to automate most of an attack chain, targeting over 20 organizations including government agencies and intelligence services, and hitting embassies and diplomatic missions in Ukraine and Europe.

AI news brief: Meta launched its new AI app Muse on Tuesday, and it’s quickly catching Wall Street’s attention, with its agentic AI push also becoming a hot topic on X. According to fresh data from market intelligence firm Sensor Tower, Muse has racked up over 83,000 downloads on US iOS, and the app is currently US-only. That download volume has already pushed Muse to the No. 2 spot on the App Store charts.

AI news brief: Alibaba had the largest haul at 151 million conversations, concentrated between May and July, peaking at nearly 3 million requests in a single day, spread across 3,500 accounts, but because they shared the same fixed prompt to trick the model into revealing its chain of thought, Anthropic determined this was a single large-scale distillation campaign run by Alibaba to train its Qwen model family — the largest such campaign the company has observed.

AI news brief: OpenAI has temporarily closed new sign-ups for its $200/month Pro plan after demand for its new flagship model Astra surged, citing strained system infrastructure. The news was announced on X by Thibault (Tibo) Sottiaux, OpenAI’s product lead for Codex, ChatGPT, and other core products, who said the Pro plan puts the most strain on the system, hence the pause on new sign-ups at that tier.

AI News Brief: Amid a surge in run-related searches, Google Search recently rolled out three AI-powered features to help runners train. This year, searches for ‘running clubs,’ ‘how to choose running shoes,’ and ‘marathon training plans’ have hit all-time highs. First, Search’s AI Mode can help build personalized training plans — through the Canvas tool in the ‘+’ menu, users can ask the system to build a schedule based on their running experience and location…

AI News Brief: OpenAI has launched ‘ChatGPT for Financial Services,’ targeting finance industry clients by combining built-in financial data with the GPT-6 Astra model for research analysis, financial modeling, and producing client-ready report materials. This lets finance professionals handle everything from data lookup to model calculations to final deliverables on one platform, skipping the back-and-forth between multiple tools…

AI News Flash: GPT-Live-1 is OpenAI’s new voice API model that lets developers build natural full-duplex voice conversation experiences in their own apps — meaning users and the AI can talk at the same time and interrupt each other anytime, without waiting for one side to finish before responding, much closer to the rhythm of real human conversation. This update strengthens the model’s instruction-following ability, so the AI can more accurately understand and execute users’ specific requests during voice interactions.

AI News Brief: OpenAI has partnered with the U.S. General Services Administration (GSA) to announce expanded AI access for government agencies, offering preferential terms to eligible federal, state, local, and tribal government bodies. This includes $0 license fees — meaning agencies can get usage licenses for free — a 50% discount on standard usage fees, and enhanced cybersecurity defense support.