AI-NATIVE INVENTION ENGINEERING A Human-Led, Evidence-Gated Architecture for Research, System Design, and Early Validaton

Description AI-Native Invention Engineering — Part II: Open Problem Infrastructure 2030–2035 extends the evidence-gated methodology introduced in Part I from individual technological projects to distributed research communities and global technological commons. The whitepaper asks which public-interest technological foundations should become independently verifiable, reproducible, interoperable, and locally adaptable by 2030–2035. It proposes twelve interconnected research missions: independent AI assurance; machine-verifiable science; AI infrastructure for low-resource languages; safe and open embodied-AI systems; local natural-risk forecasting; accessible water monitoring; local energy resilience; open agrifood intelligence; public-health early-warning infrastructure; modular assistive technology; pre-waste resource circulation; secure and vendor-independent digital public infrastructure. Each mission is expressed through an operational contract containing a problem definition, baseline, boundary, open technological layer, measurable targets for 2030 and 2035, metrics, falsifiers, localisation requirements, and safety restrictions. The paper also introduces shared infrastructure for distributed research: an Open Problem Registry, machine-readable Evidence Packages, a Negative Results Commons, reproducible open research hardware, risk-tiered openness, and a federated network of local experimental nodes. Its central objective is to move beyond isolated repositories toward Open Source Problem Solving: a system in which independent researchers can discover a well-defined problem, reproduce the existing baseline, examine previous failures, conduct a bounded experiment, and return verifiable evidence to a shared knowledge infrastructure. The twelve missions constitute an authorial research agenda rather than an official programme of the United Nations, UNESCO, OECD, or any other institution. The document does not replace professional engineering, scientific institutions, regulation, certification, or domain expertise.