Hi there!

I’m Saumya Pareek, a PhD student in the Human-Computer Interaction Group at The University of Melbourne.

My work is supported by the Google PhD Fellowship and the Len Stevens Scholarship.

My research centres around human-computer interaction and human-AI decision-making. More specifically, my research focuses on understanding the various factors that influence human decision-makers' trust and reliance on AI systems, and on designing AI explanations that better align with human reasoning. I'm also interested in examining how users' trust in AI systems can be (rightfully) recovered once lost. Ultimately, I wish to design techniques that support users in calibrating their trust to match the true capabilities of the automation they are interacting with.

I am supervised by A/Prof Jorge Goncalves and Prof Eduardo Velloso.

Contact me at: spareek [at] student.unimelb.edu.au

Find an up to date CV here

Publications

2025

• S. Pareek, S. Schömbs, E. Velloso, J. Goncalves. "It’s Not the AI’s Fault Because It Relies Purely on Data": How Causal Attributions of AI Decisions Shape Trust in AI Systems. In Proceedings of the ACM SIGCHI Conference on Human Factors in Computing Systems 2025 (CHI’25), 2025. Honorable Mention Award
• N. Boonprakong, S. Pareek, B. Tag, J. Goncalves, T. Dingler. Assessing Susceptibility Factors of Cognitive Biases in News Feed Reading. In Proceedings of the ACM SIGCHI Conference on Human Factors in Computing Systems 2025 (CHI’25), 2025.
• S. Gunasekara, S. Pareek, R. M. Kelly, J. Goncalves. The Influence of Content Modality on Perceptions of Online Misinformation. In Proceedings of the ACM SIGCHI Conference on Human Factors in Computing Systems 2025 (CHI’25), 2025.
• J. Govers, S. Pareek, E. Velloso, J. Goncalves. Feeds of Distrust: Investigating How AI-Powered News Chatbots Shape User Trust and Perceptions. ACM Transactions on Interactive Intelligent Systems (TIIS), 2025.
• C. Sew, S. Pareek, J. Govers, S. Schömbs, R. M. Kelly, J. Goncalves. The Impact of Human-Likeness and Self-Disclosure on Message Acceptance in Virtual AI Influencers. In Proceedings of the ACM SIGCHI Conference on Designing Interactive Systems (DIS'25), 2025.

2024

• S. Pareek, N. van Berkel, E. Velloso, J. Goncalves. Effect of Explanation Conceptualisations on Trust in AI-assisted Credibility Assessment. In Proceedings of the ACM on Human-Computer Interaction – CSCW, 2024. TL;DR

  TL;DR: Traditional methods of explaining AI decisions to end-users are often overly technical and cognitively demanding. This study examines how a variety of explanations that better match human reasoning (consensual, expert, internal, and empirical) impact user trust in and reliance on an AI. We find that all explanation types lead to greater reliance compared to no explanations, but also that different explanations lead to significantly different levels of reliance on the same AI decision. We also find that users often struggle to discern the AI's accuracy in guiding them toward the truth, and note instances of over-reliance and automation bias. These results provide implications for how the framing of AI explanations can moderate reliance on AI systems, impacting users’ decision-making processes.

• S. Pareek, E. Velloso, J. Goncalves. Trust Development and Repair in AI-Assisted Decision-Making during Complementary Expertise. The ACM Conference on Fairness, Accountability, and Transparency (FAccT'24), 2024. TL;DR

  TL;DR: This study investigates trust development, erosion, and repair in human-AI decision-making when human expertise complements AI capabilities. It establishes complementary expertise through tasks where users are experts for some facets, while non-experts for others. We explore various trust repair strategies, such as Apology, Denial, Promise, and Model Update. The findings show that trust established through the AI’s perceived accuracy in familiar tasks carries over to trust exhibited in unfamiliar tasks. We also find that Model Update is the most effective strategy for restoring trust, highlighting that repair strategies need not have an affective component to be effective. These insights highlight the importance of how AI systems address errors and the implications for designing AI interactions that build and maintain user trust, when warranted.

• S. Pareek, J. Goncalves. Peer–supplied Credibility Indicators as an Online Misinformation Intervention. International Journal of Human-Computer Studies (IJHCS), 2024. TL;DR

  TL;DR: This study explores the effectiveness of peer-supplied credibility labels in reducing belief in misinformation on social media. Using a 2x2 design, it investigates how the strength of social ties and political agreement between peers influence belief change. We find that credibility disputes by co-partisans (peers with similar political beliefs) significantly reduce belief in misinformation, even more than tie strength alone. However, disputes from hyper-partisan members of the opposite party can backfire, reinforcing belief in misinformation. This study underscores the potential of leveraging social media's existing social infrastructure to implement peer-supplied interventions as a means to effectively combat misinformation, and provides granular insights into their effectiveness.

• S. Schömbs, S. Pareek, J. Goncalves, W. Johal. Robot-Assisted Decision-Making: Unveiling the Role of Uncertainty Visualisation and Embodiment. In Proceedings of the ACM SIGCHI Conference on Human Factors in Computing Systems 2024 (CHI’24), 2024. TL;DR

  TL;DR: This study investigates how a robot’s certainty or confidence, visualised through either a graphical user interface (GUI) or its embodied behaviour, affects human decision-making in a high-stakes collaborative task. Results show that participants' trust and reliance on the robot vary depending on how uncertainty is conveyed. When the robot was 100% confident, participants relied more heavily on the robot when confidence was conveyed through the GUI, than with embodied cues. However, at lower confidence levels, embodied visualisations more effectively conveyed uncertainty, resulting in a more cautious approach by participants. These insights reveal how the design of uncertainty communication can shape human-agent collaboration.

• S. Wadinambiarachchi, R.M. Kelly, S. Pareek, Q. Zhou, E. Velloso. The Effects of Generative AI on Design Fixation and Divergent Thinking. In Proceedings of the ACM SIGCHI Conference on Human Factors in Computing Systems 2024 (CHI’24), 2024. TL;DR

  TL;DR: This study investigates how using generative AI as an inspiration tool impacts design fixation and divergent thinking during a visual ideation task. Our results demonstrate that viewing AI-generated imagery for inspiration leads to higher fixation, resulting in participants producing fewer, less varied, and less original ideas. Furthermore, participants often become fixated on the AI-generated images, a phenomenon called "fixation displacement." These findings have implications for designing AI systems that better support creativity and ideation without limiting exploration.

• S. Pareek, N. van Berkel, E. Velloso, J. Goncalves. How Different Explanation Conceptualisations Influence Trust in AI-Assisted Credibility Assessments. In Adjunct Proceedings of the ACM SIGCHI Conference on Human Factors in Computing Systems (CHI’24 EA), 2024.

Education

• PhD

  Dec 2022 - Present

  Human-Computer Interaction
  University of Melbourne, Australia

• Master of Information Technology (with distinction)

  Feb 2021 – Nov 2022

  Specialising in Human-Computer Interaction
  University of Melbourne, Australia
  91.11% (Dean's Honours List)


• Bachelor of Science, Computer Science (Honours)

  Jul 2017 – Aug 2020

  Mathematics minor
  University of Delhi, India
  9.65/10.0 GPA (Gold Medal)

Experience

• Head Tutor
  INFO90007: Social Computing
  INFO30005: Advanced Interface Prototyping

  
  Jul 2022 – Present
  Feb 2024 – Jun 2024

  University of Melbourne, Australia
  
  • Prepared and conducted in-person workshops for graduate and undergraduates.
  • Managed student email queries and coordinated with other tutors.
  • Evaluated student assignments and projects.

• Academic Tutor
  INFO10003: Fundamentals of Interaction Design
  INFO30005: Advanced Interface Prototyping

  
  Jul 2023 - Nov 2023
  Feb 2023 – Jun 2023

  University of Melbourne, Australia
  
  • Prepared, designed, and delivered tool demos and workshops.
  • Managed student email queries and evaluated student assignments and projects.

• UX Designer • Vocational Placement

  Feb 2022 – June 2022

  Telstra (Melbourne, Australia)
  
  • Designed a Digital Twin concept-prototype of Melbourne’s entire tram network which goal-seeks to find route optimisations, reduces transit times, tracks schedule adherence and predicts future states
  • Translated gathered insights into design requirements, designing several iterations of UI prototypes
  • View the case study here

• Data Analytics Intern

  May 2018 – Aug 2018

  International Centre for Genetic Engineering and Biotechnology (New Delhi, India)
  
  • Analysed genomic data of Malaria-causing parasites to identify disease-causing sequences
  • Created data analysis pipelines and conducted both exploratory and confirmatory research

Master's Projects

Awards & Scholarships

• Google Ph.D. Fellowship in Human-Computer Interaction

  2024

• Len Stevens Scholarship

  2024

  University of Melbourne, Australia

• Best Presentation Award, Doctoral Consortium

  2024

  University of Melbourne, Australia

• People’s Choice Award, Graduate Research Showcase

  2024

  University of Melbourne, Australia
  Awarded for the most engaging and impactful poster presentation, as voted by attendees.

• Google Conference Scholarship

  2024

  Google, Australia & New Zealand
  Provided to present a paper at a top-tier computer science conference.

• E. W. P. Kernot Memorial Bursary

  2023

  University of Melbourne, Australia
  Awarded in recognition of outstanding results in the Master of Information Technology programme.

• Dean's Honours List

  2022

  University of Melbourne, Australia
  Awarded in recognition of outstanding academic performance, representing the top 5% of students enrolled in the University's Master of Information Technology programme.

• Australian Government Research Training Program Scholarship (ARTPS)

  2022

  University of Melbourne, Australia

• Melbourne Graduate Scholarship

  2022

  University of Melbourne, Australia
  Awarded in recognition of outstanding academic achievements during first year of studies

• Melbourne School of Engineering Foundation Scholarship

  2021

  University of Melbourne, Australia
  Awarded in recognition of exceptional academic merit

• Melbourne School of Engineering Scholarship

  2021

  University of Melbourne, Australia
  Awarded based on prior academic merit to study at the university

• Honors Graduate, 2nd Rank, B.Sc. Computer Science

  2020

  University of Delhi, India