
AI Industry Research
Research directions, library, standards, briefings, thematic reports and global digital governance.
Six research directions
- Intelligent economy
- Large models & agents
- AI safety & governance
- AI industry ecosystem
- AI talent
- Industry digital transformation
Eleven industrial AI disciplines
Organized around a 'one foundation, one beam, four pillars, two industries' research layout spanning chips, architectures, software, systems and sector applications.
Embodied intelligence
Robots that perceive, decide and act in real physical environments for assembly, inspection and service.
Industrial AI chips & systems
Dedicated AI chips, accelerators and hardware-software co-design for industrial edge inference.
Novel neural architectures
Efficient network architectures, training methods and compute systems beyond today's Transformers.
Reconfigurable factory automation
Intelligent control and automation enabling rapid line reconfiguration and flexible production.
Trustworthy AI software
Safe, reliable, explainable and verifiable AI software for critical industries.
Industrial design & simulation
AI-driven industrial design, digital twins and high-fidelity simulation platforms.
Intelligent optimization
Optimization algorithms and decision systems for scheduling, energy and logistics.
Autonomous systems
Unmanned systems and smart equipment with autonomous perception, planning and control.
Multi-agent technology & platforms
Multi-agent collaboration, communication and orchestration, plus enterprise agent platforms.
Ocean sensing & communication
Intelligent sensing, underwater communication and data processing for marine environments.
Physical intelligence
AI models combining physical laws with data for scientific computing and engineering prediction.
White papers & research
White papers and reports will offer online preview and PDF download, alongside media interviews with President Zhu (e.g. International Finance News).
Among the first global AI safety standards
AIII contributes to the following standards and introduces the AISTR certification, group standards and industry drafts.
Generative AI application security testing
Security testing for generative AI applications.
LLM safety testing methods
Methods for testing large language model safety.
Large-model supply chain security
Security requirements across the model supply chain.
AI agent runtime safety testing
Runtime safety testing for AI agents.
Shanghai AI Industry Development Briefing
Co-published monthly with People's Daily Online Shanghai, available to read and download.
Sector reports
- AI in manufacturing
- AI in urban governance
- AI safety & compliance
- OPC super-individual economy
Global AI governance cooperation
Participating in UN and WDTA-related work and the DigiBridge digital capacity program.
Evaluation consulting, platform partnerships, research collaboration, events and media.
Consult us