Human Vs AI Summaries of Exploratory Data Analysis: Trust, Accuracy, And Decision Impact

References

[1] Almadani, B., Kaisar, H., Thoker, I. R., & Aliyu, F. (2025). A systematic survey of distributed decision support systems in healthcare. Systems, 13(3), 157.https://doi.org/10.3390/systems13030157

[2] Baker, A., Perov, Y., Middleton, K., Baxter, J., Mullarkey, D., Sangar, D., DoRosario, A., & Johri, S. (2020).A comparison of artificial intelligence and human doctors for the purpose of triage and diagnosis. Frontiers in Artificial Intelligence, 3, 543405.https://doi.org/10.3389/frai.2020.543405

[3] Barsalou, M. (2023). Exploring exploratory data analysis: An empirical test of run chart utility. Management Systems in Production Engineering. https://doi.org/10.2478/mspe-2023-0050

[4] Bian, X., Zhou, X., Fang, H., Liu, K., Gan, X. C., Zhao, H. N., & An, Y. (2011). Application of data visualization and numerical analysis for evaluating propagation of uncertainty. Applied Mechanics and Materials, 103, 343–347.https://doi.org/10.4028/www.scientific.net/AMM.103.343

[5] Brereton, R. G. (2021). The often-overlooked power of summary statistics in exploratory data analysis. Journal of Chemical Information and Modeling. https://doi.org/10.1021/acs.jcim.1c00244

[6] Conroy, M., Gillmann, C., Harvey, F., Mchedlidze, T., Fabrikant, S., Windhager, F., ... Betz, G. (2024). Uncertainty in humanities network visualization. Frontiers in Communication.https://doi.org/10.3389/fcomm.2023.1305137

[7] Devalla, S., & Yogi, M. K. (2023). Building trust in AI: A simplified guide to ensure software quality. Journal of Soft Computing Paradigm, 5(3), 218–231.https://doi.org/10.36548/jscp.2023.3.001

[8] Durán, J. M., & Pozzi, G. (2025). Trust and trustworthiness in AI. Philosophy & Technology, 38, 16. https://doi.org/10.1007/s13347-025-00843-2

[9] Guan, Y., Qasim, M., et al. (2022).Data intelligence and analytics: A bibliometric analysis of human–artificial intelligence in public sector decision-making effectiveness. Technological Forecasting and Social Change, 174, 121201.https://doi.org/10.1016/j.techfore.2021.121201

[10] Hassan, N., Slight, R., Bimpong, K., et al. (2024).Systematic review to understand users’ perspectives on AI-enabled decision aids to inform shared decision making. npj Digital Medicine, 7, 332.https://doi.org/10.1038/s41746-024-01326-y

[11] Hinterberger, H. (2009). Exploratory data analysis. In Encyclopedia of Database Systems. Springer. https://doi.org/10.1007/978-0-387-39940-9_1384

[12] Hongoh, V., Michel, P., Gosselin, P., et al. (2016).Multi-stakeholder decision aid for improved prioritization of the public health impact of climate-sensitive infectious diseases. International Journal of Environmental Research and Public Health, 13(4), 419.https://doi.org/10.3390/ijerph13040419

[13] Khan, M. T., Fong, C., & Tripathi, S. (2025).“Trust, but verify”: A reflexive thematic analysis of human–AI interaction. Advances in Social Sciences Research Journal, 12(11), 237–251.https://doi.org/10.14738/assrj.1211.19642

[14] Ko, D., Norton, J. W., & Daigger, G. T. (2025).Evaluation of stakeholder assessment and engagement techniques for incorporation into structured decision-making processes. Water Environment Research.https://doi.org/10.1002/wer.70187

[15] Konieczny, M., & Sporek, P. (2025).Artificial intelligence versus human analysis: Interpreting data in elderly fat reduction study. Advances in Integrative Medicine, 12(1), 13–18.https://doi.org/10.1016/j.aimed.2024.12.011

[16] Lade, S., Singh, A., Gore, A., Chaudhari, H., & Dhumal, A. (2024). Exploratory data analysis web application. International Journal for Research in Applied Science and Engineering Technology. https://doi.org/10.22214/ijraset.2024.65260

[17] Li, X., et al. (2025). The top 100 most-cited articles on large language models in medicine. Digital Health. https://doi.org/10.1177/20552076251365059

[18] Luchian, A. C., et al. (2025).From insights to trust: A review of AI-driven business analytics literature. Proceedings of the International Conference on Business Excellence, 19(1), 184–200.https://doi.org/10.2478/picbe-2025-0017

[19] Maier, T., Menold, J., & McComb, C. (2022).The relationship between performance and trust in AI in e-finance. Frontiers in Artificial Intelligence, 5, 891529.https://doi.org/10.3389/frai.2022.891529

[20] Manulak, K. B., Shaiful, Wisda, & Resyi, M. (2024). Implementation of a rule-based decision support system in determining the level of customer satisfaction with services. Ceddi Journal of Information System and Technology, 3(2), 30–37.https://doi.org/10.56134/jst.v3i2.85

[21] Mienye, D., et al. (2025). Large language models: An overview of foundational architectures, recent trends, and a new taxonomy. SN Applied Sciences. https://doi.org/10.1007/s42452-025-07668-w

[22] Mishra, A., et al. (2024). Use of large language models as AI tools in academic research and publishing among global clinical researchers. Scientific Reports. https://doi.org/10.1038/s41598-024-81370-6

[23] Mugaanyi, J., et al. (2024). Evaluation of large language model performance and reliability for citations and references. Journal of Medical Internet Research. https://doi.org/10.2196/52935

[24] Neef, A., & Neubert, D. (2011).Stakeholder participation in agricultural research projects: A conceptual framework for reflection and decision-making. Agriculture and Human Values, 28, 179–194.https://doi.org/10.1007/s10460-010-9272-z

[25] Padilla, L., Kay, M., & Hullman, J. (2021). Uncertainty visualization. In Wiley StatsRef: Statistics Reference Online.https://doi.org/10.1002/9781118445112.stat08296

[26] Putatunda, S., Ubrangala, D., Rama, K., & Kondapalli, R. (2019). SmartEDA: An R package for automated exploratory data analysis. Journal of Open Source Software. https://doi.org/10.21105/joss.01509

[27] Robertson, I. (2025).AI, trust and reliability. Philosophy & Technology, 38, 94.https://doi.org/10.1007/s13347-025-00924-2

[28] Starks, H., Shaw, J. L., Hiratsuka, V., Dillard, D. A., & Robinson, R. (2015).Engaging stakeholders to develop a depression management decision support tool in a tribal health system. Quality of Life Research, 24(5), 1097–1105.https://doi.org/10.1007/s11136-014-0810-9

[29] Thokala, P., & Madhavan, G. (2018).Stakeholder involvement in multi-criteria decision analysis. Cost Effectiveness and Resource Allocation, 16(Suppl 1). https://doi.org/10.1186/s12962-018-0120-0

[30] Venkatesh, J. D., Jaiswal, A., & Nanda, G. (2024).Comparing human text classification performance and explainability with large language and machine learning models using eye-tracking. Scientific Reports, 14, 14295.https://doi.org/10.1038/s41598-024-65080-7

[31] Vriens, M., Knoester, O. J., Bosmans, J. E., et al. (2025).The involvement of stakeholders in computational model development to support decision-making in public health: A scoping review. BMC Public Health, 25, 3568.https://doi.org/10.1186/s12889-025-24683-

[32] Wang, A., & Liang, B. (2023). Analysis of intelligent decision support systems and a multi-criteria framework for assessment. Journal of Enterprise and Business Intelligence, 3(4).https://doi.org/10.53759/5181/JEBI202303022

[33] Weiskopf, D. (2022). Uncertainty visualization: Concepts, methods, and applications in biological data visualization. Frontiers in Bioinformatics, 2, 793819.https://doi.org/10.3389/fbinf.2022.793819

[34] Wu, Y., Yuan, G.-X., & Ma, K.-L. (2012). Visualizing flow of uncertainty through analytical processes. IEEE Transactions on Visualization and Computer Graphics, 18(12), 2526–2535.https://doi.org/10.1109/TVCG.2012.285

[35] Zhang, J., & Goyal, S. B. (2024). AI-driven decision support system innovations to empower higher education administration. Journal of Computers, Mechanical and Management, 3(2), 35–41. https://doi.org/10.57159/gadl.jcmm.3.2.24070

@article{1716723,
      author = {Sai Lalitesh Pothukuchi},
      title = {Human Vs AI Summaries of Exploratory Data Analysis: Trust, Accuracy, And Decision Impact},
      journal = {Iconic Research And Engineering Journals},
      year = {2023},
      volume = {6},
      number = {11},
      pages = {1063-1077},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1716723.pdf},
      abstract = {The recent progress of artificial intelligence has dramatically changed the environment of data analysis, especially with the introduction of large language models (LLMs) into the process of analysis. Such systems are also becoming able to produce automated summaries of Exploratory Data Analysis (EDA) output and can give narrative explanations of statistical trends, correlations, and anomalies in datasets. With the use of AI-assisted analytics in organizations to enhance efficiency and scalability, concerns have arisen as to the relative reliability and interpretability of summaries prepared by human analysts compared to summaries prepared by AI systems. Recent studies point to the increased role of the analytical tools based on LLM in scientific research, business-level, as well as decision-support settings, both in terms of their transformative power and obstacles to the implementation of such tools in the context of a critical decision-making process (Mienye et al., 2025; Mishra et al., 2024). The paper analyzes the relative features of EDA summaries produced by human analysts and those produced by the LLM, in terms of four dimensions of evaluation, including factual accuracy, analytical completeness, expression of uncertainty, and perceived analytical trustworthiness. The construction of EDA narratives is usually based on domain knowledge, the reasoning of contexts, and interpretive judgment by human analysts, but the construction of EDA narratives based on probabilistic language generation and pattern recognition is based on massive amounts of training data by the systems that are built using the LLM. Though AI-based summaries have their benefits in terms of speed and scalability, their interpretive accuracy and the capacity to make reasonable statements about uncertainty are still on a scholarly research agenda. In addition to correctness in the analysis, the research additionally reviews the way such differences affect the way stakeholders make decisions. The summaries of EDA can be used as a formal linkage between complicated data and strategic actions by managers, policy makers, and other parties involved. As a consequence, misinterpretation or decision bias, or lack of uncertainty in summaries due to inaccuracies, incomplete interpretations, or poor communication of that uncertainty, can occur. The current body of research on AI-driven decision systems states that AI analytics integration can have a significant effect on the quality of decisions, organizational trust in automated systems, and user confidence calibration in the outputs of analytical systems (Guan et al., 2022). This article, by exploring the interaction between human analytical thinking and AI-generated narrative summaries, contributes to the wider debate on an ethical introduction of AI in the data-driven decision-making sphere. Particularly, the paper identifies the benefits and disadvantages of both methods of analysis and explains how the hybrid human-AI work processes can enhance the reliability, openness, and credibility of the EDA-based decision support. The results will be used to guide the development of stronger analytical models that will match the speed at which AI systems can compute with the ability of human analysts in their context.},
      keywords = {Exploratory Data Analysis (EDA) Summarization, Large Language Models (LLMs), Human–AI Analytical Comparison, Decision Support Systems, Analytical Trust and Reliability,  Uncertainty Communication in Data Analysis, Stakeholder Decision Impact},
      month = {May},
      doi = {https://doi.org/10.64388/IREV6I11-1716723}
  }