Tsinghua University Institute for AI Industry Research (AIR)

What AIR is

The Tsinghua University Institute for AI Industry Research (AIR) is a research institute at Tsinghua University devoted to applied work in artificial intelligence, and Zhang Yaqin leads the effort to set up the institute^[1]. Tsinghua's own account says he took up the post on 31 December 2019 as chair professor of intelligent science and was put in charge of establishing AIR^[1]. The name carries both halves of its identity: it is an artificial intelligence institute, and it is explicitly oriented toward industry rather than toward curiosity-driven research alone.

Where it came from

The person charged with building it came to Tsinghua from a long career in industrial research. Before Baidu, Zhang worked at Microsoft for nearly 16 years, serving among other roles as corporate vice president and chairman of Microsoft's Asia-Pacific R&D Group, and as managing director and chief scientist of Microsoft Research Asia^[1]. He joined Baidu as president in September 2014 and retired from the company in October 2019^[1]. The Tsinghua news article reporting his appointment and the institute's plans was published on 26 October 2020^[1]. That path — a long stretch in a corporate research laboratory, then a senior operating role at a Chinese internet company, then a university institute — is the background against which AIR's stated aims are usually read.

What it works on

According to the same article, AIR targets the fourth industrial revolution, with three key technology directions: autonomous driving, AI plus IoT, and brain-inspired intelligence^[1]. The institute's mission is framed as using AI to drive social progress and to empower industry^[1]. Its goal is framed as becoming an international, intelligent and industrialised applied research institution for the fourth industrial revolution^[1].

In the article Zhang set out how he sees the technology's place in the economy. He likened AI to the internet 20 years earlier: an industry in itself that also merges into every other industry^[1]. He named autonomous driving, the industrial internet of things, and AI for healthcare and biotech computing as areas with large potential^[1]. He also said he firmly opposes efforts to build an AI with self-awareness^[1].

Where it sits

The words used to describe the goal are worth taking apart. Applied and industrialised point at work that ends in deployed systems rather than in papers alone; international and intelligent point at hiring and collaboration across borders and at technical ambition. The three named directions share a common shape: each pairs a general AI capability with a specific industrial setting — vehicles, factories and physical infrastructure, or the biology lab — rather than treating AI as a standalone product. That is also what Zhang's image of the internet implies. If AI behaves like a general-purpose technology that dissolves into other industries, then an institute built around industry applications is one way to make the research useful, and to learn from deployment what the remaining technical problems are.

The stated refusal to pursue self-aware AI marks a boundary on the research agenda. The institute is presented as a place for building tools that serve people and industries, not for building a machine with its own goals — a distinction that has become a recurring part of public discussion about which directions AI research should and should not take.