More about HKUST
Human Agency in the Age of AI: Towards Human-AI Co-Creation with High Control and High Automation
PhD Thesis Proposal Defence
Title: "Human Agency in the Age of AI: Towards Human-AI Co-Creation with High
Control and High Automation"
by
Mr. Leixian SHEN
Abstract:
Before the advent of Generative AI (GenAI), digital content creation mainly
relied on manual authoring tools (e.g., Adobe Illustrator), where humans
acted as operators with high precision but limited efficiency. Today, models
like Gemini and Sora have fundamentally transformed this landscape, making it
effortless to generate high-quality multimodal content. However, this rapid
advancement introduces a fundamental tension: as AI becomes more powerful and
autonomous, human agency— the meaningful ownership and control over the
creative process—tends to diminish. Users risk being downgraded from creative
authors to passive spectators.
To address this challenge, this thesis investigates the central research
question: How can we scale AI automation without sacrificing human agency? We
argue that achieving high automation and high control is not a zero-sum
trade-off, but a design choice. This thesis presents a research journey that
systematically expands the boundaries of human-AI co-creation across the
content lifecycle, transitioning from concrete system explorations to
abstract theoretical modeling.
The research unfolds across four progressive works. First, we explore the
automation stage through Data Player, a system that automatically generates
data videos with narration-animation interplay. While highly efficient, users
become mere consumers, it exposes the necessity of fine-grained user input.
Second, to address this, we expand agency to the alignment stage with Data
Playwright. By introducing the "annotated narration" paradigm, we seamlessly
synchronize natural language intents with generative capabilities, shifting
the user's role from a passive consumer to an active director. Third, pushing
beyond static outputs, we expand agency to the adaptation stage via Live
Artifacts, a novel medium that encapsulates generative specifications within
visual layers to continuously adapt to user contexts, elevating the human to
an architect of living media. Finally, to provide a generalized foundation
for these systems, we step back to perform a theoretical abstraction. We
present the Interaction-Augmented Instruction (IAI) model, which formalizes
the fundamental synergy between natural language prompts and GUI
interactions.
Ultimately, this thesis demonstrates that in the age of AI, human agency is
not what we save from machines; it is what we systematically evolve and
expand alongside them.
Date: Monday, 28 April 2026
Time: 3:00pm - 5:00pm
Venue: Room 2132C
Lift 22
Committee Members: Prof. Huamin Qu (Supervisor)
Dr. Anyi Rao (Chairperson, AMC)
Dr. Arpit Narechania