Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

From Explainable to Interactive AI: A Literature Review on Current Trends in Human-AI Interaction

Muhammad Raees, Inge Meijerink, Ioanna Lykourentzou, Vassilis-Javed Khan, Konstantinos Papangelis

TL;DR

This paper addresses the gap that AI-HCI literature overemphasizes Explainable AI while underexploring active Human-Centered and Interactive AI approaches. Using a PRISMA-guided review of 54 studies across XAI, HCAI, Collaborative AI, IML, and Hybrid AI, it maps three dimensions—AI Users, AI Implementations, and AI Goals—and identifies a prevalent focus on user experience, limited end-user agency, and scarce practical, high-stakes applications. Key findings show that many works are theoretical, end-user co-design is rare, and only a minority permit active modification of AI mechanics, highlighting a need for more participatory design and real-world experimentation. The paper offers guidelines and directions to foster true interactive AI design, including participatory evaluation, to balance autonomy and user control and extend interaction beyond explanations into co-design and adaptation in diverse domains.

Abstract

AI systems are increasingly being adopted across various domains and application areas. With this surge, there is a growing research focus and societal concern for actively involving humans in developing, operating, and adopting these systems. Despite this concern, most existing literature on AI and Human-Computer Interaction (HCI) primarily focuses on explaining how AI systems operate and, at times, allowing users to contest AI decisions. Existing studies often overlook more impactful forms of user interaction with AI systems, such as giving users agency beyond contestability and enabling them to adapt and even co-design the AI's internal mechanics. In this survey, we aim to bridge this gap by reviewing the state-of-the-art in Human-Centered AI literature, the domain where AI and HCI studies converge, extending past Explainable and Contestable AI, delving into the Interactive AI and beyond. Our analysis contributes to shaping the trajectory of future Interactive AI design and advocates for a more user-centric approach that provides users with greater agency, fostering not only their understanding of AI's workings but also their active engagement in its development and evolution.

From Explainable to Interactive AI: A Literature Review on Current Trends in Human-AI Interaction

TL;DR

This paper addresses the gap that AI-HCI literature overemphasizes Explainable AI while underexploring active Human-Centered and Interactive AI approaches. Using a PRISMA-guided review of 54 studies across XAI, HCAI, Collaborative AI, IML, and Hybrid AI, it maps three dimensions—AI Users, AI Implementations, and AI Goals—and identifies a prevalent focus on user experience, limited end-user agency, and scarce practical, high-stakes applications. Key findings show that many works are theoretical, end-user co-design is rare, and only a minority permit active modification of AI mechanics, highlighting a need for more participatory design and real-world experimentation. The paper offers guidelines and directions to foster true interactive AI design, including participatory evaluation, to balance autonomy and user control and extend interaction beyond explanations into co-design and adaptation in diverse domains.

Abstract

AI systems are increasingly being adopted across various domains and application areas. With this surge, there is a growing research focus and societal concern for actively involving humans in developing, operating, and adopting these systems. Despite this concern, most existing literature on AI and Human-Computer Interaction (HCI) primarily focuses on explaining how AI systems operate and, at times, allowing users to contest AI decisions. Existing studies often overlook more impactful forms of user interaction with AI systems, such as giving users agency beyond contestability and enabling them to adapt and even co-design the AI's internal mechanics. In this survey, we aim to bridge this gap by reviewing the state-of-the-art in Human-Centered AI literature, the domain where AI and HCI studies converge, extending past Explainable and Contestable AI, delving into the Interactive AI and beyond. Our analysis contributes to shaping the trajectory of future Interactive AI design and advocates for a more user-centric approach that provides users with greater agency, fostering not only their understanding of AI's workings but also their active engagement in its development and evolution.
Paper Structure (35 sections, 11 figures, 2 tables)

This paper contains 35 sections, 11 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Explainable AI (XAI)
  4. Human-AI Beyond Explainability
  5. Methodology
  6. Inclusion and Exclusion Criteria
  7. Search and Data Collection
  8. Study Eligibility Evaluation
  9. Definitions and Terminology
  10. Interactivity and Agency
  11. Augmentation and Control
  12. Human-centered Design
  13. Analysis of Dimensions
  14. AI Users
  15. AI Implementations
  16. ...and 20 more sections

Figures (11)

  • Figure 1: Outline of the step-by-step process followed to find, review, and assess studies. We adapted the procedure from PRISMA PAGE2021178, a widely recognized set of guidelines for systematic reviews.
  • Figure 2: Number of included studies per year fulfilling inclusion criteria. The increase in human-AI publications in recent years is evident with a higher number of studies included in the survey that were published after 2018.
  • Figure 3: Categorization and overview of dimensions identified for the analysis. The major dimensions include users, implementations (implementation details, application domains, and data modalities), and goals (user experience, explainability, interaction, and active modification) of AI. Dimensions are further granulated to evaluate studies in detail.
  • Figure 4: Selected studies target diverse user groups. Most interactive studies are targeted towards expert (AI and domain) users while many studies published under collaborative and human-centered AI do not specifically identify the target user group.
  • Figure 5: A major portion of collaborative, human-centered, and hybrid AI studies still lies in theoretical proposals of intended systems. Interactive implementations (IML, IAI) are more developed in terms of the practical implementations of their solutions.
  • ...and 6 more figures