by Chris Irick, IP.com
Two national AI plans, same summer—no coincidence
July 2025 wasn’t just busy; it was highly choreographed. Washington rolled out America’s AI Action Plan, a sweeping blueprint to speed U.S. AI build-out and push allied standards abroad. Days later at WAIC in Shanghai, Beijing unveiled its Action Plan for Global Governance of AI, and within weeks followed up domestically with the “AI Plus” guideline to wire AI across its economy through 2030–2035. That near-simultaneity isn’t random: the U.S. plan explicitly calls for exporting the American AI stack and countering Chinese influence in international bodies—inviting a rapid, global response.
China’s action plan – a wolf in wolf’s clothing
Beijing’s global plan champions UN-centered governance, international standards via ITU/ISO/IEC, public-sector leadership, data sharing, green compute, and security/traceability. The plan urges building clean-power data centers and a “unified computing power standard system,” expanding open-source ecosystems, global data-sharing platforms, and even “lawful and orderly trading of training data,” while exploring “classified and hierarchical management” and “traceability management of AI services.” It sounds cooperative; as always, the devil is in the implementation—what gets disclosed, to whom, and with what remedies if IP leaks.
The seemingly innocuous overlap
Both sides of the AI supremacy battle pitch growth and safety: infrastructure build-out (compute, data centers, power), open/interop ecosystems, and standards work. The U.S. plan doubles down on open-source/open-weight model support and national AI infrastructure; China echoes open-source language and green compute/standards coordination. By all early accounts – America and China are on the same artificial page.
The devil is always in the details: the differences
This is where security, growth, and dominance will be decided.
Rule-setting venue. China steers governance through UN/ITU/ISO/IEC with a sovereignty-first framing; the U.S. says it will contest authoritarian influence in those bodies. Getting control through an international standards body is one of China’s main strategies.
Data flows. Beijing promotes global data sharing and trading of training data alongside public-sector leadership; Washington prioritizes export-control enforcement and allied alignment.
Implementation risk climate. Even Reuters flags top-down coordination risks—overcapacity, idle data centers, and state “picking winners”—as likely drags that can distort markets and create perverse incentives.
The intended consequences of China’s AI policy
For enterprises, China’s new posture raises the odds of compliance-driven disclosure and cross-border commingling. The Global AI Governance Action Plan leans on UN/ITU/ISO/IEC venues and public-sector leadership, and calls for risk testing/traceability systems and even “lawful and orderly transactions of training data.” In practice, that means more “inspection points” (security/algorithm filings, conformance tests, procurement checklists) where model internals, data lineage, or evaluation artifacts could be demanded, retained, or replicated; and new data-sharing markets that muddy provenance and complicate enforcement across jurisdictions. Beijing is also promoting a global cooperation body for AI—another channel to shape standards that could embed disclosure or localization as de-facto access conditions.
The implementation environment compounds the risk. As Reuters notes, China’s top-down industrial playbook tends to produce overcapacity, politicized standard-setting, and pressure to “move fast,” dynamics that historically correlate with leakage and copy-first behavior—especially if local hosting, “unified computing power” expectations, or public-sector audits are in the mix.
The practical effects show up as (i) trade-secret exposure via filings, audits, local operator/admin access, and third-country projects; and (ii) patent erosion as conformance tests and artifact sharing teach competitors how to design around claims or challenge provenance. Net-net: greater disclosure, harder provenance, more leakage vectors. So it is imperative that the US and Europe act as countervailing forces against China’s push.
IP.com’s stance: Safe & Responsible AI, built for global compliance
At IP.com, we balance speed with safeguards—never training on your data, with transparent, auditable pipelines that align to U.S. and allied expectations (as we outlined in last month’s America’s AI Action Plan blog). Get the full playbook in our Responsible AI white paper and see how our dual-AI architecture (Semantic Gist® + CompassAI) de-risks innovation for patent- and trade-secret-heavy teams.
Talk to us: sales@ip.com
IP.com’s stance: Safe & Responsible AI, built for global compliance
China: Action Plan for Global Governance of AI (full text; WAIC release, July 26, 2025).
China: “AI Plus” Guideline (Guo Fa [2025] No. 11)—domestic roll-out targets and governance language.
United States: America’s AI Action Plan (July 2025)—Pillars I–III, export controls, open-source support, and standards posture.
Breakingviews analysis of China’s approach—overcapacity, planning risk, and market distortions.
