Investigating How Leaders Decide on AI Innovations: Opportunities for HCI

Paper Title

Investigating How Leaders Decide on AI Innovations: Opportunities for HCI

Publication Info

• Topic area: Decision-making processes of leaders in AI innovation and opportunities for HCI intervention.
• Keywords: AI Deciders, AI innovation, Human-Computer Interaction, Responsible AI, AI literacy, ideation, organizational decision-making, software development, FOMO, AI adoption.

Background and Problem

• Problem / challenge: Organizations pursuing AI innovation often lack structured ideation processes, HCI expertise, and sufficient AI literacy, leading to high failure rates (80–95%) in AI projects. Responsible AI considerations are frequently overlooked.
• Significance: Reducing AI project failure and ensuring responsible AI adoption are critical for maximizing benefits and minimizing harms in the increasingly AI-driven business landscape.
• Motivation and related work: Prior research has focused on AI developers, UX designers, and users but has largely neglected the decision-making processes of AI Deciders. High failure rates in AI innovation highlight gaps in strategic planning, ideation, and risk assessment.

Solution

• Proposed approach: Investigating AI Deciders’ decision-making processes to identify opportunities for HCI to improve AI innovation outcomes and reduce project failures.
• Novelty:
  1. Empirical evidence on organizational decision-making in AI innovation, highlighting gaps in ideation and HCI involvement.
  2. Identification of misconceptions among AI Deciders, including overestimation of benefits and underestimation of risks.
  3. Proposal for integrating HCI best practices and literacy into AI innovation processes.
  4. Expansion of the definition of AI project success to include market perception and stock value impacts.
• Procedure and key techniques:
  • Semi-structured interviews with 25 AI Deciders across 20 organizations in diverse industries.
  • Affinity diagramming for qualitative data analysis to identify actionable insights.
  • Examination of organizational strategies for AI concept generation, decision-making processes, and reasoning about risks and benefits.

Results

• Concrete findings:
  • Organizations often innovate without ideation, adopting the first concept that emerges without generating or assessing alternatives.
  • AI Deciders frequently overestimate AI benefits and underestimate risks, driven by FOMO and AI hype.
  • Responsible AI considerations are largely absent, with misconceptions about societal impacts and risk management.
  • HCI expertise is missing in early-stage AI innovation, limiting strategic planning and risk mitigation.
• Advantage over baselines:
  • Identifies gaps in current AI innovation practices and highlights the potential of HCI to address these issues.
  • Proposes actionable insights for integrating HCI and Responsible AI considerations into organizational decision-making.
• Experiments / evaluation:
  • Interviews conducted with participants from diverse industries, focusing on retrospective accounts of AI projects.
  • Analysis of organizational practices and decision-making structures, revealing patterns and gaps.
• Limitations and future work:
  • Findings may not generalize to all organizations or leadership perspectives.
  • Limited by participants’ willingness to share information and confidentiality constraints.
  • Future work could explore broader definitions of AI project success and develop tools to integrate HCI into AI innovation processes.

Summary

This study investigates how AI Deciders reason about AI benefits and risks, revealing critical gaps in ideation, HCI involvement, and AI literacy. Organizations often pursue AI innovation without structured processes, overestimating benefits and underestimating risks due to FOMO and hype. Responsible AI considerations are largely absent, and HCI expertise is undervalued, limiting strategic planning and risk mitigation. The paper highlights opportunities for HCI to improve AI innovation outcomes and proposes integrating HCI literacy into business education. By addressing these gaps, organizations can reduce AI project failures and achieve more impactful and responsible AI adoption.