Human-AI Collaboration Reconfigures Group Regulation from Socially Shared to Hybrid Co-Regulation

Yujing Zhang; Xianghui Meng; Shihui Feng; Jionghao Lin

cs.AIApr 9, 2026

Human-AI Collaboration Reconfigures Group Regulation from Socially Shared to Hybrid Co-Regulation

Authors: Yujing Zhang, Xianghui Meng, Shihui Feng, Jionghao Lin

Organizations: The University of Hong Kong, Hong Kong SAR, China · Carnegie Mellon University, Pittsburgh, PA, USA · Monash University, Clayton, VIC, Australia

Abstract

Generative AI (GenAI) is increasingly used in collaborative learning, yet its effects on how groups regulate collaboration remain unclear. Effective collaboration depends not only on what groups discuss, but on how they jointly manage goals, participation, strategy use, monitoring, and repair through co-regulation and socially shared regulation. We compared collaborative regulation between Human-AI and Human-Human groups in a parallel-group randomised experiment with 71 university students completing the same collaborative tasks with GenAI either available or unavailable. Focusing on human discourse, we used statistical analyses to examine differences in the distribution of collaborative regulation across regulatory modes, regulatory processes, and participatory focuses. Results showed that GenAI availability shifted regulation away from predominantly socially shared forms towards more hybrid co-regulatory forms, with selective increases in directive, obstacle-oriented, and affective regulatory processes. Participatory-focus distributions, however, were broadly similar across conditions. These findings suggest that GenAI reshapes the distribution of regulatory responsibility in collaboration and offer implications for the human-centred design of AI-supported collaborative learning.

Figures & tables

Figure 2: GenAI agent interface embedded in the platform.

Intervention Behaviour	Example Prompt
Further Reasoning	“Could you explain your answer in more detail?”
Invite Examples	“Can you give an example to support your point?”
Challenge Viewpoints	“Do you think this always holds, or might there be exceptions?”
Turn-Taking Management	“Thank you Jack! Anyone else want to share their view?”
Summarising	“Let me summarise the group discussion.”
Task Transition	“How would you like to move forward from here?”

Table 1: GenAI Prompts Used to Support Collaborative Regulation

Code	Subcode	Operational definition	Example
CoRL	Obstacle Detection	Identifying an obstacle in another member’s learning or task progress.	“ You seem stuck on this step. ”
	Strategic Action Direction	Directing another member to adopt or change a strategy.	“ Jack, you can use the formula instead. ”
	Affective Support	Regulating another member’s emotional or motivational state.	“ Don’t worry, Lily! ”
SSRL	Shared Obstacle Negotiation	Jointly identifying and discussing group-level obstacles.	A : “ We’re off track. ” B : “ Yeah, we haven’t answered the question. ”
	Shared Strategic Negotiation	Jointly proposing and negotiating alternative strategies.	A : “ Can we compare both cases? ” B : “ Yeah, let’s start with the first one. ”
	Shared Action Change	Jointly agreeing on a change of action and coordinating execution.	A : “ Let’s change our method. ” B : “ Okay, let’s do that. ”

Table 2: Episode-Level Regulatory Processes

Code	Subcode	Operational definition	Example
Cognitive	Cog–Explain	Explains task concepts or rules to build shared understanding.	“ Frozen food must be returned within 40 minutes. ”
	Cog–Reason	Uses reasoning to justify choices or compare options.	“ If we exceed £50, we lose points. ”
	Cog–Evaluate	Judges the quality or fairness of an option.	“ Choice A is too harsh. ”
Metacognitive	Meta–Orient	Frames what the group should decide or focus on.	“ Let’s start with Question 1. ”
	Meta–Plan	Proposes strategies, sequencing, or next steps.	“ Everyone give your answer, then we compare. ”
	Meta–Monitor	Checks progress, time, understanding, or agreement.	“ How much time do we have left? ”

Table 3: Utterance-Level Participatory Focus

Regulation mode	Human–AI	Human–Human	t statistic (df = 22)
	Mean Prop. (SD)	Mean Prop. (SD)
CoRL	0.066 (0.08)	0.000 (0.00)	$-2.87^{**}$
SSRL	0.083 (0.11)	0.940 (0.12)	$18.29^{***}$
CoRL&SSRL	0.851 (0.16)	0.060 (0.12)	$-13.54^{***}$

Table 4: Distribution of regulation modes by condition

Regulatory process	Human–AI	Human–Human	t statistic (df = 22)
	Mean Prop. (SD)	Mean Prop. (SD)
Affective Support	0.837 (0.20)	0.00 (0.00)	$-14.62^{***}$
Obstacle Detection	0.314 (0.21)	0.183 (0.21)	$-2.20^{*}$
Strategic Action Direction	0.808 (0.21)	0.060 (0.12)	$-10.58^{***}$
Shared Obstacle Negotiation	0.284 (0.17)	0.394 (0.19)	$1.26$
Shared Strategic Negotiation	0.829 (0.18)	0.824 (0.17)	$-0.07$

Table 5: Distribution of Episode-Level Regulatory Processes by Condition

Explore similar work

Generative AI User Experience: Developing Human--AI Epistemic Partnership
Mar 25, 2026Xiaoming ZhaiHuman-AI Interaction Human-AI Collaboration
AI4STEM Education Center, University of Georgia Athens, GA, USA
Searching for Synergy in Shared Workspace Human-AI Collaboration
Jun 16, 2026Nachiket Kotalwar, Rohini Das, Carolyn RoseMulti-Agent Coordination Human-in-the-Loop AI
1Carnegie Mellon University.
From Prompts to Context: An Ontology-Driven Framework for Human-Generative AI Collaboration
May 28, 2026Ngoc Luyen Le, Marie-Hélène Abel, Bertrand LaforgeAI Accountability Data Provenance
1Gamaizer, 93340 Le Raincy, France. · 2Université de technologie de Compiègne, CNRS, Heudiasyc (Heuristics and Diagnosis of Complex Systems), CS 60319 - 60203 Compiègne2026 Cedex, France. · 3Sorbonne Université, CNRS UMR 7585, LPMHE (Laboratoire de Physique Nucléaire et des Hautes Énergies), 75252 Paris cedex 05,May France.

Mar 25, 2026J/K move · Enter open · S save

Generative AI User Experience: Developing Human--AI Epistemic Partnership

Xiaoming Zhai

AI4STEM Education Center, University of Georgia Athens, GA, USA

Human-AI Interaction Human-AI Collaboration Generative AI in Education

Open paper AskPDF arXiv

Generative AI (GenAI) has rapidly entered education, yet its user experience is often explained through adoption-oriented constructs such as usefulness, ease of use, and engagement. We argue that these constructs are no longer sufficient because systems such as ChatGPT do not merely support learning tasks but also participate in knowledge construction. Existing theories cannot explain why GenAI frequently produces experiences characterized by negotiated authority, redistributed cognition, and accountability tension. To address this gap, this paper develops the Human--AI Epistemic Partnership Theory (HAEPT), explaining the GenAI user experience as a form of epistemic partnership that features a dynamic negotiation of three interlocking contracts: epistemic, agency, and accountability. We argue that findings on trust, over-reliance, academic integrity, teacher caution, and relational interaction about GenAI can be reinterpreted as tensions within these contracts rather than as isolated issues. Instead of holding a single, stable view of GenAI, users adjust how they relate to it over time through calibration cycles. These repeated interactions account for why trust and skepticism often coexist and for how partnership modes describe recurrent configurations of human--AI collaboration across tasks. To demonstrate the usefulness of HAEPT, we applied it to analyze the UX of collaborative learning with AI speakers and AI-facilitated scientific argumentation, illustrating different contract configurations.

Abstract

Explore similar work

Generative AI User Experience: Developing Human--AI Epistemic Partnership

Searching for Synergy in Shared Workspace Human-AI Collaboration

From Prompts to Context: An Ontology-Driven Framework for Human-Generative AI Collaboration