Sequence CSCL research on human-AI knowledge construction from exploratory observation to experimental variable manipulation¶

Strategy · All strategies
Evidence · 2 claims (1 for, 1 mixed) · 1 study (1 theoretical), q1 · 0 of 1 report an effect size · 2 claims rest on one study

Description¶

The article proposes a concrete two-step research program. First, individuals and small groups should be systematically observed in their actual engagement with language-based AI tools, using conversation analysis, analysis of uptake events, think-aloud protocols during collaborative writing, and in-depth interviews. Subsequently, experimental research programs should vary independent variables such as prior knowledge, expectations of the tools, perceived trustworthiness, and alleged AI capabilities. Experimental studies with human-AI dyads or groups could compare the efficacy of particular prompts for learning and knowledge construction.

Design Implications¶

Context¶

Requirements¶

• Methodological breadth including ethnomethodological approaches, discourse processing, experimental studies with human-AI dyads or groups, cognitive modeling, and large-scale studies with simulated participants

Constraints¶

• Proposed for the CSCL and Learning Sciences community as future research; no findings from the program yet exist

Target Learners¶

• adults and students using ChatGPT and similar tools in everyday life

Target Learning Goals¶

• understanding how humans interact with generative AI and what they believe these tools can and cannot do
• human learning and knowledge construction in human-AI collaboration

Claims¶

• Argumentative dialogs between humans and generative AI can support knowledge transformation when prompted appropriately [+W]
• Human-AI collaboration requires refined group-awareness concepts for asymmetrical relationships between users and AI partners [~W]

Related Strategies¶

• Design experimental studies that systematically vary how GenAI outputs are presented, comparing full solutions against structured hints or constructive feedback guardrails.
• Advance CSCL by building Bayesian and collaborative-specific theoretical challenges: knowledge building priors, technology formalisation, and collaborative diagnostic reasoning
• Combine scripting with awareness measures to tailor collaboration scripts to the dyad

Examples¶

-

Key Sources¶

• Ulrike Cress · Joachim Kimmerle. (2023). Co-constructing knowledge with generative AI tools: Reflections from a CSCL perspective. Intern. J. Comput.-Support. Collab. Learn. https://doi.org/10.1007/s11412-023-09409-w