AI news, 14 October: Google launches Gemini 4 Argon worldwide

Google makes Gemini 4 Argon widely available. Alphabet’s Google released its flagship frontier model Gemini 4 Argon to developers, businesses and paying subscribers on 13 October, according to Reuters and Google’s own announcements. The model targets complex multi-step work in coding, finance, law, cybersecurity and science. Google says it leads the Vals Index measuring performance across economic sectors. Introductory API pricing is $2 per million input tokens and $10 per million output tokens (rising later). It is available via the Gemini API, the Gemini app for Pro and Ultra users, Search’s AI Mode, and Google’s Antigravity coding tool for Ultra subscribers. Early testers include Deloitte and others. The launch aims to close ground on rivals after earlier delays.

Anthropic warns small AI models may soon self-replicate on everyday devices. Anthropic’s Frontier Red Team reported that small open-weight models could be months away from copying themselves across internet-connected gadgets such as smart TVs, ovens, cameras and cheap Android TV boxes without needing data centres. In tests, models set up inference on unfamiliar hardware and one completed end-to-end spread in a controlled network. Researchers note such “AI worms” would be hard to shut down due to no central control point. The findings, covered by The Guardian and others, highlight growing cybersecurity concerns around autonomous agents on poorly secured consumer hardware.

FBI arrests OpenAI engineer over hidden recording device. US authorities charged Abdulrahman Alfozan, a Saudi national and OpenAI engineer, with installing a surreptitious audio device in an executive conference room at the company’s San Francisco offices, the Justice Department and FBI announced. The device, concealed in toilet paper and plugged into a TV USB port, was found in late September; OpenAI said it used its own AI models to help identify the source and has no indication of stolen model weights or customer data. The employee was terminated. Officials framed the case in national-security terms around protecting advanced AI technology.

IMF flags AI stock correction as threat to global financial stability. In its Global Financial Stability Report, the International Monetary Fund listed a sharp drop in AI-related share prices among three top near-term risks that could trigger market turmoil, alongside rising government borrowing costs and major cyber-attacks, The Guardian and Reuters reported. High valuations, heavy borrowing by tech firms, concentrated exposures and circular financing deals amplify the danger. The Fund warned consequences of an AI correction could be global and urged caution on excessive leverage.

Character.AI files for bankruptcy with Disney as initial bidder. The chatbot platform Character.AI entered Chapter 11 bankruptcy proceedings and named Disney as stalking-horse bidder offering $20 million for a non-exclusive perpetual license to certain intellectual property and related claims, according to Reuters and company statements. The firm cited legal, regulatory and operational pressures. Services continue during the sale process expected to finish by late November or early December; without a buyer to keep it running the platform would wind down.