Back to the library

Key question New

AI-processed · Author-approved

How does disclosing AI involvement affect support for employee ideas, and when is that penalty reduced?

When GenAI Gets the Credit: Task Non-Routinization Mitigates the Disclosure Penalty for GenAI-Initiated Creations
Dan Zhao, Vivien K.G. Lim, Thompson Teo · International Conference on Information Systems (ICIS) · 2026
Illustration for: How does disclosing AI involvement affect support for employee ideas, and when is that penalty reduced?

Loads Spotify's player; Spotify may set cookies. This device remembers your choice.

Key findings

  • Disclosing that generative AI initiated an idea and an employee refined it significantly reduced perceived employee creative credit compared to withholding disclosure.
  • Disclosing an employee-initiated and AI-refined process produced no creative credit penalty.
  • Reduced employee creative credit significantly decreased managerial willingness to sponsor the idea, willingness to implement it, and allocated financial investment.
  • Non-routine task settings weakened the negative disclosure penalty by highlighting the diagnostic value of human validation, contextualization, and accountability.

What it means for you

CIO / IT Executive

Configure enterprise GenAI tool interfaces by Monday morning to automatically log whether a prompt originated from the human employee or the AI system, capturing the exact workflow sequence before implementation.

IT Manager

Review all pending AI-assisted project submissions on Monday morning to ensure teams explicitly document that the human conceived the core idea and used AI solely for refinement, preventing unjustified credit penalties.

Business Strategist

Redesign innovation pipeline submission templates on Monday morning to explicitly frame upcoming creative challenges as highly non-routine, thereby highlighting the diagnostic value of human validation.

Researcher

Design your next experimental protocol on Monday morning to isolate the specific interaction effects between task non-routinization and human-initiated versus AI-initiated workflow sequences.

Policymaker

Draft an update to organizational AI transparency guidelines on Monday morning that differentiates mandatory disclosure requirements based on the workflow's point of origin rather than applying a blanket label.

Summary

This study examines how disclosing the division of labor between employees and generative AI affects managerial credit attribution and idea endorsement. Across two vignette experiments involving 388 full-time employees and 320 working professionals, the authors compare employee-initiated and AI-initiated workflows. The research evaluates how attribution of creative credit influences resource allocation and whether non-routine task contexts reduce disclosure penalties.

Problem

While organizations increasingly mandate transparency around generative AI use, disclosing AI assistance often penalizes the resulting creative work. Prior studies largely treat AI usage as a generic, all-or-nothing label, failing to distinguish who originated the idea versus who refined it. As a result, managers and leaders lack clear guidance on how specific human-AI workflows and task environments shape managerial evaluation and support.

Transcript
Host: Welcome to A.I.S. Insights — powered by Living Knowledge. I’m Anna Ivy Summers.

Expert: And I’m Alex Ian Sutherland.

Host: Today, we are looking at a study presented at the International Conference on Information Systems in 2026, titled "When GenAI Gets the Credit: Task Non-Routinization Mitigates the Disclosure Penalty for GenAI-Initiated Creations." The authors are Dan Zhao, Vivien K.G. Lim, and Thompson Teo. Alex, this study examines what happens when employees disclose using generative AI in their creative work. What is the fundamental problem here?

Expert: Many organizations are adopting transparency guidelines that require employees to disclose generative AI use. However, prior research shows an AI disclosure penalty: when decision-makers learn that AI was involved in an output, they often evaluate the work less favorably. The problem is that past studies generally treated AI use as a broad, binary label. They did not distinguish the internal division of labor, specifically whether the human or the AI originated the idea.

Host: That distinction seems critical for real workflows. How did the authors examine that division of labor?

Expert: They ran two experimental vignette studies. In Study 1, with 388 full-time employees, they tested perceived employee creative credit across three workflows: employee-only, employee-initiated with AI refinement, and GenAI-initiated with employee refinement. In Study 2, with 320 employed MBA professionals, they examined downstream managerial endorsement—measured as willingness to sponsor, willingness to implement, and financial budget allocation—while testing whether task context influences those judgments.

Host: What were the key findings when comparing who initiated the idea?

Expert: The results showed that the disclosure penalty depends directly on who initiated the process. Disclosing a GenAI-initiated, employee-refined process significantly reduced perceived employee creative credit compared to non-disclosure. However, disclosing an employee-initiated, GenAI-refined process produced no significant penalty.

Host: So the penalty is not an automatic reaction against any AI involvement, but specifically targets AI initiation. How does this credit loss translate into actual organizational decisions?

Expert: Perceived creative credit served as the key attributional bridge. In Study 2, lower creative credit significantly reduced managers' willingness to sponsor the idea, willingness to implement it, and the amount of budget allocated to it. Managers allocate resources not just to an idea, but to a person who can champion, explain, and execute it. When the employee does not receive credit for origination, perceived ownership weakens.

Host: The authors also identified a boundary condition: task non-routinization. How did that alter evaluations?

Expert: In routine tasks with standardized procedures, refining an AI draft provided few cues of distinctive employee skill, so the disclosure penalty remained severe. But under high task non-routinization—marked by market volatility, unclear positioning, and unpredictable requirements—the negative disclosure penalty was significantly smaller. In complex settings, managers recognized that contextualizing, validating, and modifying an initial AI draft requires meaningful human judgment and accountability.

Host: What are the main takeaways for organizations managing AI transparency?

Expert: First, organizations should avoid generic "AI-generated" tags in submission workflows. Provenance records should detail who framed the problem, who drafted the initial concept, and what specific validation and contextualization the employee performed. Second, leaders must be careful not to penalize AI initiation indiscriminately; doing so creates incentives for employees to conceal AI use, undermining governance. Finally, managers need to separate idea quality, employee contribution, and policy compliance into distinct evaluative questions.

Host: Understanding the exact workflow and task context ensures promising ideas are not prematurely discounted. Thank you, Alex, and thank you to our listeners for tuning into A.I.S. Insights — powered by Living Knowledge.
Artificial Intelligence & Automation Information Systems & Management Human-Computer Interaction Data Analytics & Big Data Digital Transformation & Strategy Cybersecurity & Privacy Cloud & Infrastructure E-Commerce & Digital Platforms Operations & Supply Chain Sustainability & Green IS

Generative AI, Authorship disclosure, Creative credit, Idea endorsement, Task non-routinization, Human-AI collaboration