--- title: "2026.10.06 AI news briefing — Opening of AI Festa 26 and NYC Council big tech hearing" url: https://blog.tyrano.dev/en/2026-10-06-ai-news-briefing-opening-of-ai-festa-26-and-nyc-council-big-tech-hearing lang: en site: Tech AI News - 테카이 category: "AI News" tags: ["ai news","daily","ai festa","nyc council","openai","anthropic"] published: 2026-10-06T04:33:50.095Z updated: 2026-10-06T04:34:21.863Z sources: - https://zdnet.co.kr/view/?no=20261005111000 - https://www.dailian.co.kr/news/view/1686494/%EB%B9%85%ED%85%8C%ED%81%AC-%EC%A4%91%EC%8B%AC-%ED%8C%90-%EA%B9%A8%EB%8B%A4SKT%EC%B9%B4%EC%B9%B4%EC%98%A4KT-2026 - https://en.sedaily.com/finance/2026/10/05/hyundai-nvidia-and-lg-form-ai-self-driving-alliance-to-take - https://www.cnbc.com/2026/10/05/anthropic-openai-google-meta-execs-testify-nyc-council-ai-hearing.html - https://techcrunch.com/2026/10/05/openai-will-start-watermarking-chatgpts-text-in-the-eu/ - https://www.pymnts.com/news/artificial-intelligence/2026/microsoft-meta-steer-staff-from-anthropic-claude-in-house-ai/ --- # 2026.10.06 AI news briefing — Opening of AI Festa 26 and NYC Council big tech hearing > At Seoul COEX, the largest-ever “AI Festa 26” opened, and the government-led “Everyone’s AI” consortium began its October beta. Overseas, the New York City Council put four big tech firms on the stand over AI safety, and OpenAI introduced text watermarking to comply with EU rules. --- ## 1. ‘AI Festa 26’ opens at COEX — largest ever - **Source:** [ZDNet Korea](https://zdnet.co.kr/view/?no=20261005111000) | 2026.10.05 - **Key points:** The Ministry of Science and ICT’s “AI Festa 26” will be held October 6–8 at COEX Hall C in Seoul. Under the theme “AI becomes real. Everyone’s AI,” 212 companies and 501 booths are participating, marking the largest scale to date. In Korea, LG AI Research, SK Telecom, and Upstage are joining as the three new entrants with proprietary foundation models, while from abroad Google DeepMind, OpenAI, Perplexity, and scholars including Yoshua Bengio are taking part. On the first day (the 6th), the main conference “AI Summit” will be held. - **Why it matters:** It’s a venue to gauge both the competitive landscape among three Korean proprietary foundation-model teams and the level of engagement from global big tech, all in one place. ## 2. ‘Everyone’s AI’ — SKT, KT, Kakao to launch beta in October - **Source:** [Dailian](https://www.dailian.co.kr/news/view/1686494/%EB%B9%85%ED%85%8C%ED%81%AC-%EC%A4%91%EC%8B%AC-%ED%8C%90-%EA%B9%A8%EB%8B%A4SKT%EC%B9%B4%EC%B9%B4%EC%98%A4KT-2026) | 2026.10.05 - **Key points:** Under the government-backed “Everyone’s AI” initiative, three consortia—SKT, KT, and Kakao—will run betas from September to October and aim for an official launch in December. Kakao will let users handle everything from information search to payments via familiar channels like KakaoTalk and phone calls, powered by Kanana and EXAONE. KT will offer a real-time search–based general-purpose AI and “Ieum,” a suite of 10 specialized services across finance, education, and more. SKT will provide a phone- and SMS-based action-taking AI that also covers non-smartphone users. Next year’s government budget is 2,500억원. - **Why it matters:** It’s the first large-scale experiment in which Korean telecom and platform providers, backed by government support, jump into the big tech–centered AI assistant race. ## 3. Hyundai, Nvidia, and LG form AI self-driving alliance - **Source:** [Seoul Economic Daily](https://en.sedaily.com/finance/2026/10/05/hyundai-nvidia-and-lg-form-ai-self-driving-alliance-to-take) | 2026.10.05 - **Key points:** Hyundai Motor Group is partnering with Nvidia and LG Electronics to develop a next-generation AI autonomous driving platform. It will combine Nvidia’s Drive Hyperion system with LG’s in-vehicle infotainment technology into a high-performance computing platform, with plans to mass-produce Level 2++ autonomous vehicles starting in 2028. - **Why it matters:** Analysis: With Tesla FSD entering the Korean market, this signals Korean companies’ countermove toward an “AI-defined vehicle” ecosystem. ## 4. New York City Council holds AI safety hearing with four big tech firms - **Source:** [CNBC](https://www.cnbc.com/2026/10/05/anthropic-openai-google-meta-execs-testify-nyc-council-ai-hearing.html) | 2026.10.05 - **Key points:** The New York City Council convened a committee-of-the-whole hearing with all 51 members, taking sworn testimony from executives at OpenAI, Anthropic, Google, and Meta. Chair Julie Menin secured attendance after warning of subpoenas, and the Council is considering bills that would bar deploying AI models in New York City without external validation. At the same session, whistleblowers including former Anthropic employee Jacob Coxon warned that “at the current pace, humanity is more likely to lose control over AI.” - **Why it matters:** It’s the first large-scale attempt by a U.S. local government to impose legal accountability on AI firms, and if enacted, similar legislation could spread to other cities. ## 5. OpenAI to introduce text watermarking for ChatGPT and Codex in the EU - **Source:** [TechCrunch](https://techcrunch.com/2026/10/05/openai-will-start-watermarking-chatgpts-text-in-the-eu/) | 2026.10.05 - **Key points:** To comply with the EU AI Act’s transparency rules, OpenAI will apply an invisible watermark (“textGrain”) to text generated by ChatGPT and Codex. It subtly adjusts word choices with a secret key to leave a detectable pattern. The feature will roll out automatically to EU users within weeks, and API developers can enable it immediately. However, OpenAI’s own tests found that changing just 10% of words drops detection accuracy from 92% to 66%. - **Why it matters:** Even while acknowledging current detection limits, OpenAI is prioritizing regulatory compliance. Whether this spreads to other regions and companies bears watching. ## 6. Microsoft and Meta scale back in-house use of Anthropic Claude - **Source:** [PYMNTS](https://www.pymnts.com/news/artificial-intelligence/2026/microsoft-meta-steer-staff-from-anthropic-claude-in-house-ai/) | 2026.10.05 - **Key points:** According to The Information, Microsoft and Meta are steering employees toward their own AI tools, reducing use of Anthropic’s Claude. Microsoft cut its expected annual Claude spend by one-third, shifting users to GitHub Copilot. Meta halved Claude Code users from about 60,000 to 30,000, transitioning them to in-house tools “Muse Code” and “MetaCode.” - **Why it matters:** Big tech is moving to reduce dependence on external models, a signal that could impact Anthropic’s enterprise revenue mix. --- ## Today’s insights - In Korea, government-led AI infrastructure and services (AI Festa, Everyone’s AI) surfaced simultaneously, making it a key moment to see whether public–private collaboration can deliver results. - Abroad, safety and regulation (NYC Council, EU watermarking) and supplier diversification within big tech (Microsoft, Meta) are unfolding in parallel, suggesting the AI industry is seeking a balance between growth and control.