AI in the Eyes of Users: Public Perception Survey Report on AI (Part 2)
-
This is the second installment of the AI survey series. The article explores users' perception of product intelligence, purchasing decisions, and usage experiences through three sections: 'Intelligence in Users' Eyes,' 'Smart Hardware Product Experience,' and 'Leopard Insights.'
When it comes to 'artificial intelligence,' the industry generally defines it as 'human-like intelligence.' AI experts aim for machines to 'think like humans,' 'act like humans,' 'think rationally,' and 'make rational decisions and actions.'
However, as AI moves from the lab into households as consumer products, do ordinary people perceive smart hardware products as 'intelligent' in the way experts expect?
User-perceived intelligence, as discussed here, refers to the sense of intelligence users feel when interacting with smart products, which may include perceptions of new technologies or human-like experiences.
Consumer-facing smart hardware powered by AI, as tangible entities in users' lives, delivers the strongest sense of intelligence and reflects how deeply smart products have integrated into daily routines.
Selected smart hardware products in this article:
What defines intelligence for users? This question can be broken down into three aspects:
For users, products (apps, mini-programs, hardware, etc.) are considered intelligent if they feature automatic adjustments, voice interaction, remote control, and similar capabilities. Human-like appearances, however, are less associated with perceived intelligence, especially those resembling humans closely.
This indicates that users associate intelligence more with the application of new technologies rather than physical mimicry.
Users with deeper knowledge of AI are more likely to recognize features like autonomous learning, gesture interaction, human-like emotions, and appearances as indicators of stronger intelligence.
Data shows that the features users associate with intelligence align with what they desire in products, suggesting they value and seek out these intelligent traits.
Smart hardware falls into two categories:
- AI-enhanced traditional hardware (e.g., TVs, cameras, speakers), which adds new interaction methods and expanded functionalities to core features.
- New AI-driven hardware (e.g., robots), primarily reliant on AI technology.
In the previous article, we found users acknowledge the positive impact of smart hardware on life. Here, we focus on smart home and wearable devices—products deeply intertwined with daily life—to understand users' real experiences with purchasing and using them.
AI × IoT transformation is the inevitable future for TVs. Currently, voice-interactive smart TVs dominate the market, representing early-stage AI applications. Key interaction methods include far-field voice (hands-free, no remote) and near-field voice (remote-assisted).
When buying AI TVs, users prioritize picture clarity and smooth playback, followed by voice interaction and tech-savvy features. About 55% of AI TVs are priced between ¥1,000–3,000.
Users express high satisfaction with AI TVs, particularly praising the efficiency and freedom of voice interaction.
However, since older adults form a significant TV-watching demographic, voice interaction—still limited by technical and implementation challenges—can feel burdensome for them to learn.
Smart speakers have become a gateway to smart homes. According to IDC's China Smart Home Device Quarterly Tracker, Q1 2019 saw shipments of 11.22 million units, a 787.2% year-on-year increase.
This means voice interaction and AI services are becoming part of daily life for millions of Chinese households.
When purchasing speakers, users prioritize core functionalities like sound quality and music library breadth, with smart features ranking mid-tier. Educational attributes appeal mainly to families with children. In AIoT applications, speakers are often used to control lights and outlets. About 51% of smart speakers cost under ¥200.
User satisfaction with smart speakers is high for perceived intelligence but lower for practicality. While they bring voice interaction to homes, some users report lacking a compelling reason for sustained use.
Affordable smart home security products like cameras are gaining rapid popularity among younger users for enabling smart living at low cost.
For smart cameras, high-definition and panoramic views are top priorities, while smart tracking and tech appeal rank lower. Privacy risks are also a major concern. About 71% of smart cameras cost under ¥200.
Users rate smart cameras highly for satisfaction and practicality. Monitoring focuses on children and the elderly, with pet surveillance also driving purchases among pet lovers.
In personal/home service robots, cleaning robots like robotic vacuums address household chores, making them early adopters in consumer markets.
Prospective Industry Research notes China's robotic vacuum sales hit 5.77 million units in 2018, up 42.1% year-on-year.
Cleaning performance is the top purchase factor for robotic vacuums, with climbing/obstacle avoidance less critical. Prices vary widely: 35% under ¥500, 45% over ¥1,000.
User satisfaction with robotic vacuums is relatively low, citing charging issues and subpar cleaning.
The home service robot industry is in its growth phase, expanding rapidly with high potential. Future years will see fast market growth, moderate new capacity, and high entry barriers (CIC).
For companion robots, natural voice interaction, functionality, and content richness are key considerations. Most market offerings are educational robots for kids, with 45% of top sellers priced under ¥500.
Users report positive experiences with companion robots, which overlap functionally with smart displays.
From a user perspective, AI TVs and smart cameras score high for intelligence and utility, while robotic vacuums—requiring minimal interaction—are perceived as less intelligent and practical.
Users associate intelligence with AI applications like auto-adjustment, voice interaction, and remote control.
Technically, these features simulate, extend, or augment human intelligence, making them inherently human-like.
But for users, convenience matters more than human likeness. The 'uncanny valley' effect also shows that overly human-like robots can trigger discomfort.
When purchasing, core product functionalities dominate decisions—cleaning performance for vacuums, picture quality for TVs.
Combining these insights, we recommend focusing R&D and marketing on AI's practical benefits over human-like aesthetics.
Price-wise, smart living is accessible: ¥100 buys a speaker or camera, ¥500 gets a vacuum and educational robot, and ¥1,000–2,000 secures a voice-controlled AI TV.
New interaction modes and controls in smart products do enhance convenience, as confirmed by this survey—though criticisms remain.
For example, in a home setting, who should the product listen to during multi-person voice interactions? How should users handle faults in increasingly integrated products?
Such complaints reveal that current interaction designs for smart hardware still focus on the product rather than the user.
The emergence of AI has shifted interaction design from a single product experience to one that must consider accurate, timely, and personalized services for users in various scenarios and emergencies, while also maintaining the product's 'persona' and avoiding user offense.
Facing such significant changes, providing more elegant and convenient solutions to convey a sense of 'intelligence' to users may be a new direction that requires special attention in designing smart products.
AI in the Eyes of Ordinary People: Public AI Perception Research Report (Part 1)