Explainable & Ethical AI
Explainable Personality-Aware AI for Adaptive Academic Support
Develop and evaluate an explainable NLP framework that uses carefully bounded personality-related signals from text to support adaptive, ethical academic guidance without making automated high-stakes decisions.
- 0
- Progress
- 2
- Researchers
About this project
This project will investigate whether personality-aware text representations can improve academic support tools while preserving transparency, privacy, fairness and human oversight. The team will create a reproducible benchmark using appropriate existing or consented text data, compare interpretable baselines with graph and neural approaches, and build a decision-support prototype for researchers and educators. Planned outputs include a peer-reviewed paper, reusable code, evaluation protocols, and dataset and model documentation. The work will not perform clinical diagnosis or automate consequential academic decisions.
Objectives
- Define a responsible, human-centred framework for personality-aware academic support
- Build a reproducible benchmark with clear data provenance and consent boundaries
- Compare interpretable baselines with graph-based and neural NLP models
- Evaluate predictive performance, calibration, fairness, robustness and explanation quality
- Develop and assess a transparent decision-support prototype
- Publish reusable methods, code and reporting guidance
Research priorities
- Explainable and ethical AI
- Personality computing and natural language processing
- Fairness, privacy and human oversight
- Cross-cultural validity and robustness
- Reproducible research and open documentation
- Human-computer interaction in education
Sustainable Development Goals
- SDG 4 · Quality education
- SDG 9 · Industry, innovation and infrastructure
- SDG 10 · Reduced inequalities
Activities
- Protocol development and focused literature review
- Dataset identification, provenance review and documentation
- Baseline and advanced model implementation
- Explainability, fairness, calibration and robustness evaluation
- Human-centred prototype design and formative evaluation
- Paper, code, model card and dataset documentation preparation
How to take part
Researchers can contribute to literature synthesis, ethical and data-governance design, NLP and machine-learning implementation, graph modelling, evaluation, explainability, HCI studies, reproducible software, or scholarly writing. Contributions will be documented and authorship will follow substantive contribution.
Skills required
- Python
- natural language processing
- machine learning
- explainable AI
- evaluation
- research methods
- scholarly writing
Roles needed
- NLP and ML researcher
- Responsible AI researcher
- HCI researcher
- Data and evaluation researcher
- Academic writing and review
Who should apply
Open to postgraduate students, early-career researchers and experienced researchers who can demonstrate relevant skills or a clear capacity to contribute. All participants must follow ethical data handling, reproducible workflows and contribution-based authorship.
Expected contribution
Collaborators should take ownership of agreed tasks, document methods and decisions, participate in regular research meetings, use reproducible version-controlled workflows, respect privacy and research ethics, and contribute substantively to outputs.
