Guided By AI: Navigating Trust, Bias, and Data Exploration in AI-Guided Visual Analytics
2404.14521
0
0
š
Abstract
The increasing integration of artificial intelligence (AI) in visual analytics (VA) tools raises vital questions about the behavior of users, their trust, and the potential of induced biases when provided with guidance during data exploration. We present an experiment where participants engaged in a visual data exploration task while receiving intelligent suggestions supplemented with four different transparency levels. We also modulated the difficulty of the task (easy or hard) to simulate a more tedious scenario for the analyst. Our results indicate that participants were more inclined to accept suggestions when completing a more difficult task despite the AI's lower suggestion accuracy. Moreover, the levels of transparency tested in this study did not significantly affect suggestion usage or subjective trust ratings of the participants. Additionally, we observed that participants who utilized suggestions throughout the task explored a greater quantity and diversity of data points. We discuss these findings and the implications of this research for improving the design and effectiveness of AI-guided VA tools.
Create account to get full access
Overview
- The paper explores the impact of AI-powered suggestions on user behavior and trust during a visual data exploration task.
- Participants completed the task with varying levels of transparency around the AI system's suggestions.
- The task difficulty was also manipulated to simulate a more tedious scenario for the analysts.
- The study found that users were more inclined to accept AI suggestions when the task was more challenging, despite lower suggestion accuracy.
- Transparency levels did not significantly affect suggestion usage or trust in the AI system.
- Users who utilized the AI suggestions explored a greater quantity and diversity of data points.
Plain English Explanation
The paper investigates how people interact with and trust artificial intelligence (AI) systems that provide suggestions during visual data exploration. The researchers conducted an experiment where participants had to explore and analyze some data, and the AI system provided them with suggestions on what to look at next.
The key things the researchers looked at were:
- How much the participants relied on the AI's suggestions, especially when the task was more difficult.
- How much the participants trusted the AI system, and whether that was affected by how much information (transparency) the system provided about its suggestions.
- How the AI suggestions influenced the participants' exploration of the data - did they look at more or different things than those who didn't use the suggestions?
The main findings were:
- Participants were more likely to follow the AI's suggestions when the task was harder, even though the suggestions were less accurate in those cases.
- The level of transparency about how the AI system worked didn't really affect how much people trusted it or used its suggestions.
- People who used the AI suggestions tended to explore the data more thoroughly and look at a greater variety of information.
The researchers discuss how these results can help improve the design and effectiveness of AI-powered tools for visual data exploration and decision-making.
Technical Explanation
The researchers conducted an experiment to investigate the impact of AI-powered suggestions on user behavior and trust during a visual data exploration task. Participants were asked to complete a data exploration task, with the task difficulty manipulated (easy or hard) to simulate a more tedious scenario for the analyst.
During the task, participants received intelligent suggestions from an AI system, with the level of transparency around those suggestions varied across four conditions: no transparency, low transparency, medium transparency, and high transparency. The researchers measured how often participants accepted the AI's suggestions, their subjective trust in the system, and the quantity and diversity of data points they explored.
The results showed that participants were more inclined to accept the AI's suggestions when completing the more difficult task, despite the system's lower suggestion accuracy in that condition. Interestingly, the levels of transparency tested did not significantly affect suggestion usage or trust ratings. Additionally, the researchers observed that participants who utilized the AI suggestions throughout the task explored a greater quantity and diversity of data points.
The findings from this study have implications for enhancing trust in machine learning models and improving the design of AI-guided visual analytics tools. The researchers discuss how the transparency levels tested may not have been sufficient to impact user trust, and suggest further research to understand the specific transparency needs of users in these interactive visual analytics scenarios.
Critical Analysis
The study provides valuable insights into how users interact with and trust AI-powered suggestions during visual data exploration tasks. However, there are a few limitations and areas for further research that could be addressed:
-
Transparency Levels: The researchers acknowledged that the transparency levels tested may not have been granular enough to significantly impact user trust. Further research could explore a wider range of transparency mechanisms and examine their specific effects on trust and reliance on AI suggestions.
-
Individual Differences: The study did not account for individual differences in data analysis skills, cognitive abilities, or prior experiences with AI systems. These factors could influence how users respond to and trust the AI's guidance, and should be explored in future studies.
-
Long-Term Interactions: This was a single-session experiment. Examining how user trust and reliance on AI suggestions evolve over extended interactions could provide valuable insights for the design of AI-guided visual analytics tools.
-
Real-World Applicability: While the experimental setup aimed to simulate a more tedious data exploration scenario, the generalizability of the findings to real-world, complex visual analytics tasks remains to be assessed.
Overall, this study contributes important findings to the understanding of how users engage with and trust AI-powered suggestions in visual data exploration. The insights can inform the design of more effective and trustworthy AI-guided systems for data analysis and decision-making.
Conclusion
This experiment investigated the impact of AI-powered suggestions on user behavior and trust during a visual data exploration task. The key findings suggest that users are more inclined to rely on AI suggestions when the task is more challenging, despite lower suggestion accuracy. Surprisingly, the transparency level of the AI system did not significantly affect suggestion usage or subjective trust. Additionally, participants who utilized the AI suggestions explored a greater quantity and diversity of data points.
These results have important implications for the design and effectiveness of AI-guided visual analytics tools. By understanding how users interact with and trust AI-powered guidance, researchers and practitioners can develop more trustworthy and effective systems to support data exploration, analysis, and decision-making. Further research is needed to explore the nuances of transparency, individual differences, and long-term interactions to refine our understanding of this critical topic.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Papers
HAIChart: Human and AI Paired Visualization System
Yupeng Xie, Yuyu Luo, Guoliang Li, Nan Tang
0
0
The growing importance of data visualization in business intelligence and data science emphasizes the need for tools that can efficiently generate meaningful visualizations from large datasets. Existing tools fall into two main categories: human-powered tools (e.g., Tableau and PowerBI), which require intensive expert involvement, and AI-powered automated tools (e.g., Draco and Table2Charts), which often fall short of guessing specific user needs. In this paper, we aim to achieve the best of both worlds. Our key idea is to initially auto-generate a set of high-quality visualizations to minimize manual effort, then refine this process iteratively with user feedback to more closely align with their needs. To this end, we present HAIChart, a reinforcement learning-based framework designed to iteratively recommend good visualizations for a given dataset by incorporating user feedback. Specifically, we propose a Monte Carlo Graph Search-based visualization generation algorithm paired with a composite reward function to efficiently explore the visualization space and automatically generate good visualizations. We devise a visualization hints mechanism to actively incorporate user feedback, thus progressively refining the visualization generation module. We further prove that the top-k visualization hints selection problem is NP-hard and design an efficient algorithm. We conduct both quantitative evaluations and user studies, showing that HAIChart significantly outperforms state-of-the-art human-powered tools (21% better at Recall and 1.8 times faster) and AI-powered automatic tools (25.1% and 14.9% better in terms of Hit@3 and R10@30, respectively).
6/18/2024
š¤
Harder, Better, Faster, Stronger: Interactive Visualization for Human-Centered AI Tools
Md Naimul Hoque, Sungbok Shin, Niklas Elmqvist
0
0
Human-centered AI (HCAI), rather than replacing the human, puts the human user in the driver's seat of so-called human-centered AI-infused tools (HCAI tools): interactive software tools that amplify, augment, empower, and enhance human performance using AI models; often novel generative or foundation AI ones. In this paper, we discuss how interactive visualization can be a key enabling technology for creating such human-centered AI tools. Visualization has already been shown to be a fundamental component in explainable AI models, and coupling this with data-driven, semantic, and unified interaction feedback loops will enable a human-centered approach to integrating AI models in the loop with human users. We present several examples of our past and current work on such HCAI tools, including for creative writing, temporal prediction, and user experience analysis. We then draw parallels between these tools to suggest common themes on how interactive visualization can support the design of future HCAI tools.
4/3/2024
š¤æ
The Drawback of Insight: Detailed Explanations Can Reduce Agreement with XAI
Sabid Bin Habib Pias, Alicia Freel, Timothy Trammel, Taslima Akter, Donald Williamson, Apu Kapadia
0
0
With the emergence of Artificial Intelligence (AI)-based decision-making, explanations help increase new technology adoption through enhanced trust and reliability. However, our experimental study challenges the notion that every user universally values explanations. We argue that the agreement with AI suggestions, whether accompanied by explanations or not, is influenced by individual differences in personality traits and the users' comfort with technology. We found that people with higher neuroticism and lower technological comfort showed more agreement with the recommendations without explanations. As more users become exposed to eXplainable AI (XAI) and AI-based systems, we argue that the XAI design should not provide explanations for users with high neuroticism and low technology comfort. Prioritizing user personalities in XAI systems will help users become better collaborators of AI systems.
5/1/2024
š¤
Overcoming Anchoring Bias: The Potential of AI and XAI-based Decision Support
Felix Haag, Carlo Stingl, Katrin Zerfass, Konstantin Hopf, Thorsten Staake
0
0
Information systems (IS) are frequently designed to leverage the negative effect of anchoring bias to influence individuals' decision-making (e.g., by manipulating purchase decisions). Recent advances in Artificial Intelligence (AI) and the explanations of its decisions through explainable AI (XAI) have opened new opportunities for mitigating biased decisions. So far, the potential of these technological advances to overcome anchoring bias remains widely unclear. To this end, we conducted two online experiments with a total of N=390 participants in the context of purchase decisions to examine the impact of AI and XAI-based decision support on anchoring bias. Our results show that AI alone and its combination with XAI help to mitigate the negative effect of anchoring bias. Ultimately, our findings have implications for the design of AI and XAI-based decision support and IS to overcome cognitive biases.
5/9/2024