WAILabs Psych Corner: How AI Interactional Systems Shape Psychology - And Why Gender Matters

From Psychology to AI Ethics: A Journey in Understanding Human-Technology InteractioN

By Dr. Auxane Boch

As a researcher at the intersection of psychology, human-computer interaction (HCI), and AI ethics, my work has long been driven by a simple yet urgent question: How do interactive and interactional systems - not just as tools, but as entities and experience creators - shape our emotions, behaviours, interactions with others and even our sense of self? This question has led me from the lecture halls of Université Savoie Mont Blanc, where I first studied psychology, to the cutting-edge labs of the Institute for Ethics in Artificial Intelligence (IEAI) at the Technical University of Munich (TUM). Today, as an Associate Research Director and WAI Labs Psychology Lead, I’m focused on unpacking the psychological impacts of AI interactional systems (from video games to social robots) with a critical lens on diversity.

My academic path reflects this interdisciplinary focus. After completing my BSc in Psychology in France, I pursued an MSc in Cyberpsychology at Dún Laoghaire’s Institute of Art, Design and Technology in Ireland, where I explored how video games can help us reflect on our perceptions and attitudes towards AI. My doctoral research in Social Sciences (Dr. rer. pol.) at TUM deepened my expertise in the societal implications of AI, particularly in understanding and defining how psychology can be integrated into the understanding of these technologies’ impacts and, thus, their governance. Now, deepening this line of thought, I launched the Psychology Impact Assessment for Interactional Systems (PSAIS) project, aiming to develop a multicultural framework to evaluate the psychological effects of interactional systems and ensure that technologies are designed with human values, equity, and trust at their core.

The WAI Labs Psych Corner: A Hub for Applied AI

I’m thrilled to be affiliated with WAI Labs, an initiative that embodies the mission of making AI more inclusive, ethical, and human-centred. The Psych Corner - a new arm of WAI Labs - is dedicated to exploring how interactive systems such as chatbots, social robots, and AI companions influence our wellbeing, identity, cognitions, behaviours and social connections. Our goal is to bridge the gap between technical innovation and psychological insight, ensuring that AI development is grounded in evidence-based design.

At the heart of this effort is a systematic review we are conducting: “Gender-specific psychological impacts of interactional systems: a systematic review.” This project, registered with PROSPERO, is the first of its kind to comprehensively synthesise empirical evidence on how AI systems affect psychology - and how these effects differ by gender. Why gender? Because the impacts of AI are not one-size-fits-all. From voice assistants reinforcing stereotypes to social robots influencing emotional regulation, design cues and societal norms can amplify or mitigate psychological outcomes in ways we’re only beginning to understand.

The Systematic Review: Why It Matters

Our review follows the PRISMA methodology, a gold standard for systematic reviews, to ensure rigour and reproducibility. We’re screening thousands of studies across five major databases - Scopus, PubMed, IEEE Xplore, ACM Digital Library, and Google Scholar - focusing on five categories of AI interactional systems:

1. Chatbots & LLM Agents (e.g., conversational AI, voice assistants)

2. AI Companions & Relational Agents (e.g., digital humans, virtual agents)

3. Social Robots & Embodied Agents (e.g., humanoid robots, HRI systems)

4. Recommender Systems & Algorithmic Feeds (e.g., personalised content, ranking algorithms)

5. AI-Enabled XR/VR/AR & Adaptive Games (e.g., immersive simulations, AI-driven gaming)

Our inclusion criteria are strict: studies must be peer-reviewed, empirical (quantitative, qualitative, or mixed-methods), and published in the last decade (2016-2026). They must also measure psychological outcomes, such as anxiety, social connectedness, self-efficacy, or cognitive load, and analyse them through a gender lens, one way or another. This means we’re excluding non-peer-reviewed work and studies without an accessible full text.

Key Innovations in Our Approach

• Gender as a Moderator: We’re not just looking at average effects; we’re examining how gender differences manifest across outcomes. For example, do women experience greater cognitive load when using voice assistants? Are men more susceptible to identity shifts in AI companion interactions?

• Interdisciplinary Integration: By combining insights from psychology, HCI, HRI, and UX research, we’re building an actionable evidence base for developers, policymakers, and researchers.

• Methodological Rigour: Each study is screened by three independent reviewers using Rayyan, a collaborative tool for systematic reviews. Disagreements are resolved through discussion, with final adjudication by me as the lead researcher.

Our search strategy is equally thorough. We’ve designed five separate searches per database, each targeting a specific AI system type combined with psychological outcomes and gender/sex terms. For example, in Scopus, we’ve identified 106 relevant entries for chatbots alone, with similar traction in IEEE Xplore (514 entries) and PubMed (925 entries). The sheer volume of raw data (pre-screening) underscores the need for synthesis - and the potential for transformative insights.

The Team Behind the Work

This review is a collaborative effort with two exceptional co-researchers:

• Dr. Kathleen Belhassein, researcher at Institut PPRIME (CNRS)

• Dr. Seren Yenikent, founder of ai-mind.solutions

  
  

What We Hope to Achieve

By Q4 2026, we aim to submit a peer-reviewed paper outlining our findings, alongside an evidence map and design recommendations for industry and policymakers. Our goals are threefold:

1. Map the Landscape: Identify which psychological outcomes (e.g., stress, autonomy, emotional well-being) are most affected by AI systems and how these effects vary by gender.

2. Uncover Mechanisms: Determine which system features (e.g., persona design, adaptivity, modality) drive beneficial or harmful outcomes.

3. Translate to Action: Develop testable design recommendations that developers can implement to mitigate risks and enhance positive impacts.

For example, if we find that female users report higher anxiety when interacting with male-voiced chatbots, we might recommend gender-neutral voice options as a default. Or if social robots are shown to improve social connectedness for lonely older adults, we could advocate for their integration into elder care, with safeguards to prevent over-reliance.

Final Thought: AI for Humans, by Humans

This work isn’t just academic - it’s a call to action. Our findings will equip product managers with evidence to design inclusive AI technology, researchers with gaps to explore, companies with practical frameworks to build ethically, and policymakers with insights to regulate responsibly. By grounding AI in psychology, we ensure that technology serves everyone - not just the average user, but people of all genders, ages, and cultures. Reach out to wailabs@womeninai.co if you’re a researcher, developer, or policymaker interested in collaborating or learning more.

Acknowledgements

Generative-AI tool use: 

Generative AI tools were used solely for language editing (Grammarly) and text refinement/rephrasing for clarity (Mistral AI). All ideas, analyses, and conclusions are the author’s own.

Disclaimer

This blog is a contribution of expertise from our volunteers. It is not a reflection of any opinion or roadmap of their employers. All blogs from WAI Labs go through a review and/or editing as needed, and are vetted for veracity. For questions or comments, write to wailabs@womeninai.co .