Imagining machine vision: Four visual registers from the Chinese AI industry

Area of Science:

• Computer Science
• Sociology of Technology

Background:

• Machine vision is a key artificial intelligence (AI) application, dominating China's AI market.
• Technologies like facial recognition and automated driving are integral to surveillance and social governance.
• Sociotechnical imaginaries shape the definition and development of emerging technologies.

Purpose of the Study:

• To analyze how leading Chinese companies represent machine vision technologies.
• To understand the role of sociotechnical imaginaries in shaping these representations.
• To identify the visual registers used to articulate these imaginaries.

Main Methods:

• Qualitative multimodal analysis of corporate websites of major Chinese machine vision players.
• Examination of visual and textual content to identify recurring themes and representations.
• Application of the concept of sociotechnical imaginaries to interpret findings.

Main Results:

• Identification of a cohesive sociotechnical imaginary of machine vision in China.
• Discovery of four distinct visual registers: computational abstraction, human-machine coordination, smooth everyday, and dashboard realism.
• These registers contribute to a narrow and opaque portrayal of machine vision.

Conclusions:

• Chinese tech companies utilize specific sociotechnical imaginaries to represent machine vision.
• The identified visual registers serve to articulate global ambitions and competitiveness.
• These representations may obscure the complexities and implications of machine vision technologies.