AI News, 15 November 2026

Sunday 15 November

Major AI companies implement the agreed common technical standard for training opt-outs by the end of 2027. — 42% say yes

AI-generated summary OpenAI, Google, Microsoft, and Anthropic reached an agreement on an artificial intelligence training opt-out standard targeting a rollout by the end of 2027. A panel of six frontier models leans no on whether the standard will be implemented across primary consumer and developer platforms by that deadline.

Johnson Controls and Armada begin production at Dublin AI infrastructure plant

Source-page capture accompanying Johnson Controls and Armada begin production at Dublin AI infrastructure plant

AI-generated summary Johnson Controls and Armada announced the start of production at Galleon Forge Two, a 90,000-square-foot manufacturing plant in Dublin, Ireland, to build modular computing infrastructure for sovereign artificial intelligence workloads. The Dublin facility employs roughly 700 workers and produced its first Leviathan modular data-centre unit for the Fossefall project in Norway, with the partners targeting at least 500 MW of production capacity in 2027 across multiple sites.

Earlier:

South Korean Police Launch Investigation Into Wide Scale Exam Fraud After AI Cheating Admission

South Korean Police Launch Investigation Into Wide Scale Exam Fraud After AI Cheating Admission

AI-generated summary South Korean police have launched a wider sweep for technology-assisted fraud in professional certification tests following a cheating admission involving generative artificial intelligence on the national tax accountant exam. The candidate who ranked first confessed to utilizing generative AI and electronic devices to obtain answers during the examinations before turning himself in to authorities.

Earlier:

China calls for AI cooperation with the US ahead of Shenzhen Apec summit

AI-generated summary Chinese ambassador to the United States Xie Feng stated that China and the United States should cooperate on artificial intelligence. Xie Feng promoted the World Artificial Intelligence Cooperation Organisation as a venue for regional prosperity and development.

Earlier:

Perplexity Decisions API Adds pplx-decider-v1.3, flash-v1.3, and Text-Span Matching via locate

Figure 1 Perplexity Decisions API (November 2026) model specs, aliases, and official pricing

AI-generated summary Perplexity announced two updated decision models, pplx-decider-v1.3 and pplx-decider-flash-v1.3, for its Decisions API alongside a new locate question type. Both new models support a 256K-token context window while maintaining previous pricing rates of $0.02 and $0.01 per 1M input tokens.

Also covered byGitSeonjeap
Earlier:

Baseten publishes DeepSeek engine, reports 2,060 tokens per second on eight B300s

AI-generated summary Baseten published an inference engine for the DeepSeek V4.1 Flash model, reporting a speed of 2,060 output tokens per second on eight Nvidia B300 GPUs. Teams with compatible hardware can run the DeepSeek weights on the engine, which was written in Rust and CUDA and released under the Elastic License 2.0.

Earlier: