AI News, 16 November 2026

Monday 16 November

(2nd LD) PM calls for considering disabling Wi-Fi at CSAT test sites to prevent AI cheating

AI-generated summary Prime Minister Han Seong-sook instructed officials to consider disabling public Wi-Fi signals at all test sites of this week's college entrance exam to prevent artificial intelligence cheating. The measures follow a confession by a top scorer on a certified tax accountant exam who admitted to using a smartphone, miniature camera, and generative artificial intelligence to cheat.

NVIDIA Platforms Gain Traction in Korean Physical AI and Agentic Workloads

AI-generated summary NVIDIA platforms gained traction among Korean enterprises and startups deploying robotics, edge artificial intelligence, and agentic workflows at AI Day Seoul. Over two thousand attendees participated in the event to explore hardware and software stacks including Isaac, DGX, and OpenShell.

Also covered byTokenTimesaiunderstanding.orgTechPowerUp
Earlier:

I tested the Halliday G2 AI smart glasses and they turned me into a walking 'cheat sheet' — but everyone could tell

AI-generated summary The Halliday G2 AI smart glasses feature dual micro LED engines and waveguide optics to deliver green dot matrix style text alongside a Halliday AI assistant. The device includes Meeting Flow for transcribing and summarizing conversations, a Cheat Sheet teleprompter feature, and a Translate tool for over 45 languages.

How to Use the github-security Extension in the GitHub Copilot App (Public Preview)

DigIntoAI · AI model news, analysis and practical guides

AI-generated summary According to official documentation, developers can install the github-security plugin in the GitHub Copilot app to unlock a dedicated Security canvas for triaging alerts and reviewing local changes. The extension is in public preview and remains subject to change, requiring an active Copilot plan, the GitHub Copilot app, and a signed-in GitHub account.

Earlier:

Ontario Appeal Court Keeps OpenAI Copyright Suit in Canada, Where US Fair Use Does Not Apply

OpenAI Logo

AI-generated summary Ontario's Court of Appeal ruled that OpenAI must defend a copyright lawsuit brought by five Canadian news organizations in an Ontario courtroom. The court rejected the argument that the dispute belongs in a U.S. court, noting that the U.S. fair use doctrine does not apply and Canadian law lacks a text-and-data-mining exception.

Also covered bySI Agent News
Earlier:

Ten Minutes to Diagnosis: AI System Slashes Sleep Apnea Testing Time While Outperforming Existing Models by 18%

Ten Minutes to Diagnosis: AI System Slashes Sleep Apnea Testing Time While Outperforming Existing Models by 18%

AI-generated summary Researchers developed an artificial intelligence framework called the Long-Range Interdependencies Aware framework that predicts obstructive sleep apnea severity from ten minutes of physiological data. The study was validated across over thirty-five thousand subjects and achieved accuracy rates surpassing existing models by up to eighteen percent across multiple datasets.

Meta Muse Spaces Explained: Talk, Build and Remember — What Meta's November Update Really Changes

Meta Muse Spaces voice assistant on a smartphone

AI-generated summary Meta launched Muse Spaces on 6 November 2026 to let users turn a single sentence into a small custom app inside the Muse assistant. Meta states that spaces can pull updates from connected third-party apps including Gmail, Google Calendar, OpenTable and Notion.

Earlier: