What Is Hot in CHI

Wei LiAAAI 2015 · DOI

Introduction

The ACM Conference on Human Factors in Computing Systems (CHI) represents the pinnacle of Human-Computer Interaction (HCI) research. Given its multidisciplinary nature—spanning psychology, design, sociology, and engineering—the sheer volume of work (over 2,000 submissions in 2014) can be overwhelming for researchers in adjacent fields like Artificial Intelligence. This paper synthesizes the "hottest" trends from CHI 2014, identifying key areas where AI can solve real-world interaction challenges.

Core Research Themes

The synthesis identifies three primary pillars of current HCI research: the integration of digital and physical environments, the evolution of personal fabrication, and the rise of human-computer intelligence.

1. The Digital-Physical Continuum

A significant shift is occurring from purely screen-based interfaces to those that inhabit the physical world. This is manifested through:

  • Tangible User Interfaces (TUIs): Leveraging human tactile abilities by giving digital information physical form.
  • Wearables and Activity Tracking: Utilizing diverse data streams from body-worn sensors to estimate health metrics and cognitive workload.
  • Predictive Control: Moving beyond manual interaction toward systems that model user patterns (e.g., smart HVAC systems) to automate environments.

2. The Future of Making: Personal Fabrication

Personal fabrication is transitioning from "shape prototyping" to the creation of functional electromechanical devices. CHI research now explores multi-material 3D printing, enabling the fabrication of complex items like speakers in a single process. For the AI community, the challenge lies in developing "function-based" design tools that allow non-experts to capture physical objects and transform them into functional digital designs effortlessly.

3. Human-Computer Intelligence

As computers still lack general conceptual intelligence, researchers are exploring "Collective Computation" and "Friendsourcing." Unlike traditional crowdsourcing, friendsourcing leverages social capital and social network interactions. This requires sophisticated AI models to calculate the "social cost" of human attention and incentivize participation through gamification or social obligation.

Key Insights for AI Researchers

Several findings from CHI 2014 offer profound implications for machine learning and algorithmic design:

  • The Paradox of Success: In a study of the Nest thermostat, researchers found that as predictive accuracy increased, user engagement decreased. This suggests that for smart interfaces, "low engagement" is often a metric of high performance, challenging the traditional UI goal of maximizing user interaction time.
  • Tangible vs. Augmented Reality: While Augmented Reality (AR) overlays digital data on the world, there remains a persistent demand for tangible interfaces. Humans have evolved to manipulate physical materials; AI-driven systems must account for this preference for physical graspability over pixel-based overlays.
  • Scaling Human Expertise: The "Frenzy" system demonstrated how massive organizational tasks—such as scheduling a conference with thousands of papers—can be treated as a global constraint satisfaction problem. By decomposing the task, a mix of expert committees and crowdsourced input can solve complex logistical challenges that neither could handle alone.

Methodological Evolution

Machine Learning has transitioned from a niche methodology to a standard component in HCI research. It is now ubiquitously used for:

  • Gesture and activity recognition.
  • User behavior classification.
  • Constraint satisfaction in collaborative systems.

Future Frontiers and Limitations

Despite these advancements, several gaps remain where AI research is uniquely positioned to contribute:

  • Generalizing Function-Based Design: Current tools for capturing and modifying physical objects are highly domain-specific. There is a critical need for general-purpose AI tools that lower the barrier to entry for 3D modeling.
  • Closing the Intelligence Gap: Future work must focus on more seamless human-computer collaboration, finding better ways to structure tasks so that AI can effectively augment human conceptual intelligence.

Conclusion

The trends at CHI 2014 highlight a move toward more invisible, physical, and collaborative computing. For AI researchers, the "hottest" opportunities lie in modeling social costs, automating physical design, and embracing the paradox where the best interface is the one the user needs to interact with the least.

Summary updated

BibTeX
@inproceedings{li2015what,
  author = {Wei Li},
  title = {{What Is Hot in CHI}},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year = {2015},
  doi = {10.1609/aaai.v29i1.9345},
}
Figure