- The Shift
- Posts
- China’s Global AI Plan
China’s Global AI Plan
Plus, 🤩Most Viral and Interesting AI Video Formats, Sam Altman Warns: ChatGPT Therapy Chats Aren’t Legally Private, and more!
Hello there! Ready to dive into another upgrading, Mind-boggling, and Value-filled Shift?
Today we have:
🌏 China’s Global AI Plan: Full-Spectrum Push for Open, Inclusive Governance
🤩Most Viral and Interesting AI Video Formats
🧠 Sam Altman Warns: ChatGPT Therapy Chats Aren’t Legally Private
🔨Tools and Shifts you cannot miss
🌏 China’s Global AI Plan: Full-Spectrum Push for Open, Inclusive Governance
China just unveiled a sweeping AI action plan at WAIC 2025, emphasizing joint R&D, open-source ecosystems, and governance led by the United Nations. Released just days after the U.S. plan, China’s approach positions itself as the global AI partner of choice for developing nations.
The Shift:
1. Joint R&D, Infrastructure & AI for the Global South - China is calling for cross-border research labs, data-sharing platforms, and technical capacity-building across nations, especially in the global South. It advocates open-source frameworks, shared corpora, and unified standards to lower the barrier to innovation and prevent duplicative waste.
2. UN-Led Governance & Open Collaboration - Beijing proposes a multilateral AI governance system under the United Nations, including a new global AI cooperation body. Premier Li explicitly warned that AI must not become an “exclusive game” dominated by a few nations or firms. The plan urges consensus on global norms, ethics, risk controls, and equitable access through forums like the Global Digital Compact.
3. Standards, Safety, and Public-Sector Deployment - The action plan calls for AI security testing systems, traceability, and classification-based governance to prevent misuse. It emphasizes developing high-quality, diverse datasets while protecting privacy, and promotes energy-efficient AI through low-power chips and water-saving infrastructure.
China isn’t just building models; it’s building a governance framework, infrastructure blueprint, and values system that appeals to emerging nations left out of the Western AI race. By aligning with the UN and offering real resources like computing, education, and open-source tools, Beijing is courting the global majority.
🤩Most Viral and Interesting AI Video Formats
Via Justine Moore on X
What's blowing up in AI video this week?
I curated the most viral & interesting formats across platforms - and recreated them to share workflows.
The list and how to make them 👇
— Justine Moore (@venturetwins)
2:32 PM • Jul 27, 2025
🧠 Sam Altman Warns: ChatGPT Therapy Chats Aren’t Legally Private
Many users treat ChatGPT like a digital therapist, but legally, it's nothing close. In a recent podcast with Theo Von, OpenAI CEO Sam Altman made it clear: there’s no doctor-patient confidentiality with AI, and that could open the door to serious privacy risks.
The Shift:
1. No Legal Shield for Emotional Chats - Altman explained that ChatGPT is used as a life coach or therapist by millions, especially young users. But unlike real therapists or lawyers, those conversations aren’t protected by legal privilege. If subpoenaed, OpenAI is obligated to hand over chat histories, no matter how sensitive.
2. Ongoing Legal Risks Are Mounting - In the NYT lawsuit, OpenAI is already fighting a court order that could force it to store and produce user chats globally (excluding enterprise). Altman called this demand “an overreach” and warned it may set a dangerous precedent.
3. Lack of Clarity Stalls Trust and Adoption - Altman acknowledged that privacy fears are stopping many from fully using ChatGPT, and that hesitation is justified. Without legal clarity, users are left exposed in gray areas that traditional platforms wouldn’t dare tread.
Until AI conversations earn the same legal protections as human ones, user trust remains fragile. The more ChatGPT acts like a therapist, the more urgent it becomes to lock down privacy by law, not just policy. Otherwise, the very users who need AI support the most may be the first to walk away.
🔨AI Tools for the Shift
💼 LiGo for LinkedIn – Get authentic LinkedIn posts, comments, and advanced analytics. Built for content creators and B2B pros.
🏠 Montessori Activities at Home – Turn everyday items into enriching learning experiences. Inspired by Montessori, powered by creativity.
🎵 Singify AI Vocal Remover – Split any song into vocals and instrumentals instantly. Ideal for karaoke, remixes, or covers.
🚗 Specd – Professional luxury vehicle valuations with real-time market data. Built for buyers, sellers, and auto industry experts.
📩 Privy – Turn Email & SMS into your best sales channel. Built to convert visitors into customers with popups, texts, and emails.
🚀Quick Shifts
🧠 Microsoft may be prepping for GPT-5. A New Copilot code references a “Smart Mode” powered by GPT-5, offering faster or deeper thinking, just as OpenAI’s August release window approaches.
🤖 DOGE built an AI to review 200000 federal rules, aiming to cut half by Trump’s first year back, already deployed at HUD and CFPB. The White House says no plan is officially approved yet.
🧠 Meta names Shengjia Zhao as Chief Scientist of its new Superintelligence Labs, tapping the former OpenAI researcher behind GPT-4 and o1 to lead AI reasoning breakthroughs alongside Alexandr Wang.
📈 AI referrals to websites hit 1.13B in June 2025, up 357% YoY, driven mostly by ChatGPT, which accounts for over 80% of traffic. News sites saw a 770% surge, with Yahoo, Reuters, and The Guardian leading.
Reply