Writing

Notes on AI, productization, and the questions worth asking.

Op-eds, white papers, and peer-reviewed research. The writing extends the practice — formalizing what advisory engagements treat as common sense, and putting structural arguments where they can be challenged.

The writing covers what the advisory work covers, in a register that is more permanent and more public. Most of what passes for "AI commentary" is reactive. A new model lands, a piece responds. A regulation passes, a piece responds. The cycle rewards speed over substance and rewards consensus over precision.

I write more slowly than that. The pieces below — peer-reviewed papers, applied white papers, the occasional newspaper op-ed — are the ones I have wanted to be on the record for, in environments where being wrong has consequences for the person reading. The through-line across all of them is the same as the practice: what does responsible, deployable AI actually look like in the rooms where mistakes are expensive?

Newer pieces will appear here as they publish. The list below is the corpus as it stands today — chronological, categorized, cross-linked to the venue or repository where the full text lives.

Op-Eds & Commentary

Public commentary, where it earns its place.

Pieces written for newspaper, magazine, and general intellectual audiences. The category is small by design — public commentary should be selective. More to come.

  • 2024

    Why do we self-medicate?

    BusinessDay Newspaper · Nigeria · February 4, 2024

    An op-ed on the structural conditions that drive self-medication in Nigerian households, and what that argues about how AI systems should treat the resulting health information practices.

    Read →
White Papers

Applied research released for public use.

Technical pieces written for practitioners, policymakers, and the people responsible for deploying AI in production environments. Released through UMBC's Center for Applied AI and Training Centers.

  • 2024

    A Survey of Deepfake Detection Technologies

    UMBC Center for Applied AI · Research release

    A survey of the state of deepfake detection — what works, what does not, and the structural reasons detection systems are losing ground to generation systems. Written for practitioners and policymakers.

    Read →
  • 2024

    Plagiarism and Deepfakes: The Challenge of Originality in the Age of AI

    UMBC Training Centers · Center for Applied AI · Research release

    A piece on what AI-generated content does to the concept of originality — and to the institutions (academia, journalism, publishing) that depend on it. Written for educators, editors, and the practitioners building the next generation of authentication infrastructure.

    Read →
Peer-Reviewed Papers

The full academic record.

Eight peer-reviewed papers — published, in press, under review, and in preparation — across IEEE, ACM, JMIR, JAMIA, and PMC venues. The Research page treats these as the through-line of the dissertation argument; the listing here is comprehensive.

  • 2025

    Real-Time Detection of Online Health Misinformation Using an Integrated Knowledge-Graph–LLM Approach

    Clark, O. & Joshi, K. P. · IEEE ICDH · World Congress on Services · Helsinki

    Risk-aware narrative classification combining knowledge-graph reasoning with LLM inference. IEEE ICDH Best Student Paper Award.

    ResearchGate →
  • 2025

    Evaluating Causal AI Techniques for Health Misinformation Detection

    Clark, O. & Joshi, K. P. · CARD Workshop · IEEE PerCom · Washington, DC

    Causal inference methods evaluated against narrative-aware approaches; the comparative groundwork for the Risk Irrelevance Principle.

    IEEE Xplore →
  • 2025

    Security Compliance for Smart Manufacturing Using Knowledge-Graph-Based Digital Twin

    IEEE BigData · 2025

    An applied paper extending knowledge-graph-based reasoning into the security-compliance domain for industrial digital-twin systems. Outside the dissertation through-line; published as a separate research strand.

    Recently presented · IEEE Xplore forthcoming
  • 2024

    Global Relevance of Online Health Information Sources: A Case Study of Experiences and Perceptions of Nigerians

    Clark, O., Joshi, K. P., & Reynolds, T. · AMIA Annual Symposium · San Francisco

    The empirical study that grounds the equity argument. Surveys how Nigerians evaluate, trust, and act on online health information across cultural contexts.

    PMC →
  • 2024

    Exploring the Impact of Increased Health Information Accessibility in Cyberspace on Trust and Self-Care Practices

    Clark, O., Reynolds, T., Ugwuabonyi, E., & Joshi, K. P. · ACM SaT-CPS Workshop

    The earliest formal articulation of Online Health Safety as a distinct problem space between patient safety and misinformation detection.

    UMBC Ebiquity →
  • 2026

    AI-Driven Reconstruction of Online Health Narratives: Overcoming Narrative Blindness Through Theory-Grounded Agent-Action-Outcome Modeling

    Journal of Medical Internet Research (JMIR) · Under review · second R&R

    The full Agent–Action–Outcome narrative reconstruction methodology. Formalizes Narrative Blindness and the modelling response.

    JMIR Preprint →
  • 2026

    The Online Health Safety Gap: Quantifying the Invisible Health Risk of Factually Accurate Peer-to-Peer Health Content

    Clark, O., Ahmed, Z., & Joshi, K. P. · IEEE Open Access · Under review

    The empirical paper behind the 68.8% Online Health Safety Gap finding. Quantifies the harm potential carried by factually accurate health content that consensus-trained systems cannot see.

    Under peer review
  • 2026

    Risk-Aware Online Health Narratives Classification Using a Two-Dimensional Neuro-Symbolic Approach

    IEEE International Conference on Digital Health (ICDH) 2026 · Submitted April 2026

    Extends the risk-aware classification framework into a two-dimensional neuro-symbolic implementation; submitted for peer review at IEEE ICDH 2026.

    Under peer review
  • 2026

    VERITAS: A Neuro-Symbolic Approach to Quantifying Epistemic Divergence and Harm Potential in Online Health Narratives

    Clark, O., Joshi, K. P., & Joshi, A. · Journal of the American Medical Informatics Association (JAMIA)

    The integrated framework paper. Brings the two constructs, the four-category classification, and the empirical findings into a single methodological statement.

    Due for submission
Going Further

For syndication, op-ed pitches, or research collaboration — let's talk.

Begin a Conversation See the Research →