ExplAI Workshop

---

ExplAI-2026 seeks cutting-edge research on explainable, transparent, and trustworthy artificial intelligence systems that advance the United Nations’ Sustainable Development Goals (SDGs). We welcome interdisciplinary work that fuses robust AI methodology with ethical principles, governance frameworks, and demonstrable social impact. The workshop provides an open forum for exchanging ideas, forging collaborations, and charting a path toward responsible, human-centred innovation.

Building on the first edition, whose proceedings appeared as CEUR-WS Vol-4141, the 2026 edition places particular emphasis on explainability for foundation models and large language models, on the operational requirements that emerging AI regulation and standards – notably the EU AI Act, ISO/IEC 42001, and the NIST AI Risk Management Framework – place on transparency and risk management, and on evidence that links AI systems to measurable progress against specific SDG indicators. Submissions are invited from academia, industry, and the public sector; submission is open to all researchers working on the topics below.

---

We invite original papers on theoretical and practical advances in transparent, trustworthy, and ethical AI for decision-making aligned with the UN Sustainable Development Goals. Topics include, but are not limited to, the four themes below.

Explainable AI (XAI)

–    Post-hoc explanation methods and intrinsically interpretable models

–    Concept-based, counterfactual, and example-based explanations

–    Explainability for foundation models, large language models, and multimodal systems

–    Metrics, benchmarks, and reproducible protocols for evaluating explanation quality

–    Human-centred, user-adaptive, and interactive explanation interfaces

AI ethics, governance, and compliance

–    Fairness auditing, bias detection, and bias mitigation

–    Accountability, traceability, and algorithmic transparency by design

–    Privacy-preserving, federated, and data-minimising learning

–    Alignment with emerging regulation and trustworthy-AI standards

–    Risk, conformity, and ethical impact assessment methodologies

Trustworthy and robust AI

–    Uncertainty quantification, calibration, and selective prediction

–    Robustness, adversarial resilience, and operational safety

–    Verification, validation, and certification of learning-based systems

–    Human-in-the-loop decision support and human–AI teaming

–    Reliability of AI in safety-critical and high-stakes domains

AI for the Sustainable Development Goals

–    Health and well-being; education and lifelong learning (SDG 3, SDG 4)

–    Climate action, clean energy, and sustainable cities (SDG 7, SDG 11, SDG 13)

–    Decent work, inclusive growth, gender equality, and reduced inequalities (SDG 5, SDG 8, SDG 10)

–    Green AI: energy-efficient models and sustainable computing (SDG 9, SDG 12)

–    Peace, justice, and strong institutions: AI in public administration, the rule of law, and democratic oversight (SDG 16)

–    Methods for measuring and reporting AI contributions to SDG indicators

–    Paper type. Regular papers of 10–15 pages in English, including figures, tables, and references. Page counts follow the CEUR-WS.org definition of a standard page (about 2,500 characters).

–    Template. Papers must use the one-column CEURART template. CEUR-WS.org provides LaTeX and LibreOffice (ODT) versions only; both already contain the mandatory sections described below.

–    LaTeX on Overleaf (recommended): CEURART template

–    LaTeX archive: CEURART.zip [TBA]

–    LibreOffice: CEUR-Template-1col.odt [TBA]

–    CEUR-WS.org publishes no official Word template. Authors who cannot use LaTeX or LibreOffice may request a DOCX conversion from the organisers, but remain responsible for matching the CEURART layout exactly.

–    Peer review. All submissions are reviewed double-blind by at least three members of the international programme committee. Reviewers assess originality, technical soundness, significance, relevance to the workshop scope, clarity of presentation, and reproducibility. Where reviewers disagree, the programme committee chairs invite a further review and moderate a discussion before deciding. Final decisions rest with the programme committee chairs and are based solely on the quality and relevance of the paper. Programme committee members do not review submissions with which they have a conflict of interest.

–    Submission system. Microsoft CMT [TBA] – select ExplAI-2026 in the conference list.

–    Originality. Submissions must be original and must not have been published elsewhere or be under review elsewhere. CEUR-WS.org does not publish previously published papers, nor papers rejected by another programme committee of the same event. All submissions are screened for textual overlap with published work; submissions found to contain plagiarised or undisclosed duplicated material are rejected without review.

–    Authorship. At most five authors per paper. In the camera-ready version, author names must be given in Latin characters, with meaningful affiliations and institutional e-mail addresses.

–    Presentation. ExplAI-2026 is a single-day, on-site event at one physical location; virtual track is possible. At least one author of each accepted paper must attend the event and present their work. Papers that are not presented will not be included in the proceedings submitted to CEUR-WS.org.

Every accepted paper must satisfy the current CEUR-WS.org submission rules. Please read them at ceur-ws.org/HOWTOSUBMIT.html before preparing the camera-ready version.

Volume composition

CEUR-WS.org constrains the composition of a published volume: no single institution may contribute more than 33% of the papers, counting any authorship position and not only first authorship, and no member of the organising committee may author or co-author more than one paper. A volume must also contain at least six papers and at least 40 pages excluding frontmatter. Where these limits bind, the programme committee chairs may be unable to include an otherwise acceptable paper in the proceedings; such papers may still be presented at the workshop.

Declaration on Generative AI (mandatory)

Since 1 June 2025 every paper must contain a section headed Declaration on Generative AI, placed before the references. Authors must list every generative-AI tool or service used and describe its specific contribution using the CEUR-WS GenAI Usage Taxonomy. Generating scientific content, insights, or conclusions with generative AI is not permitted, and AI tools must never be listed as authors. If no such tool was used, the section must say so explicitly. See the CEUR-WS GenAI policy and the GenAI Usage Taxonomy.

Example of declaration: “During the preparation of this work, the author(s) used ChatGPT, Grammarly in order to: Grammar and spelling check, Paraphrase and reword. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the publication’s content.”

Adapt the wording to the tools you actually used; do not copy the example verbatim. If no generative-AI tool was used, the section must still be present and say so explicitly, for example: “The author(s) have not employed any generative-AI tools in the preparation of this work.”

Licence and copyright notice

The volume and its papers are published under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). Authors retain copyright. The first page of every camera-ready paper must carry the following footer, exactly as prescribed by CEUR-WS.org:

“Copyright © 2026 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).”

In the CEURART template this notice is produced automatically by the \copyrightyear and \copyrightclause commands; do not typeset it by hand. The first-page footer must also identify the event. Set the event line exactly as follows:

\conference{ExplAI-2026: 2nd Workshop on Advanced AI in Explainability and Ethics for the Sustainable Development Goals, November 13, 2026, Khmelnytskyi, Ukraine}

Signed author agreement

Each accepted paper must be accompanied by a CEUR-WS.org open-access author agreement. One agreement per paper is required, signed by the corresponding author on behalf of all co-authors; the title and author list on the form must match the camera-ready paper exactly. Complete the form by hand, sign it with an ink pen on paper – electronic and touchscreen signatures are not accepted – and return a scan (PDF or JPEG) together with the camera-ready paper. CEUR-WS.org revises these forms periodically: download the current version at the time of signing rather than reusing an earlier copy. Use the form that matches your paper:

–    Author agreement (NTP) [TBA] – the standard case: the paper and any accompanying material contain no copyrighted third-party content.

–    Author agreement (TP) [TBA] – required if the paper reuses copyrighted third-party material. A copy of the rights holder’s permission must be attached to the signed agreement.

Author metadata and ORCID

In the camera-ready version, provide meaningful affiliations and institutional e-mail addresses in the paper footer on page 1, following the CEURART instructions. ORCID identifiers are optional in the CEURART template but are strongly recommended for every author; register free of charge at orcid.org. CEUR-WS.org may ask the editors for a DBLP footprint, so authors are encouraged to keep their DBLP record accurate and consistently transliterated.

Camera-ready compliance check

Before the volume is assembled, the organisers check every camera-ready paper against the requirements above: CEURART layout, the Declaration on Generative AI, the copyright notice and event line, author metadata, and the signed author agreement. Papers that remain non-compliant cannot be included in the proceedings.

Work involving human participants, personal data, or sensitive applications must state the ethical approvals obtained, the consent procedures followed, and the data-protection measures applied. Authors should state the provenance and licence of every dataset used, and are encouraged to include a short data- and code-availability statement. Reporting of computational cost and energy use is welcome, and is expected for submissions under the Green AI theme.

The following is guidance from the organisers rather than a set of formal requirements. Observing it will improve the reception of your paper and ease compliance with CEUR-WS.org.

–    Give the paper a specific, informative title, normally no longer than about 13 words.

–    Write in clear, proofread English; have a fluent reader check the text before submission.

–    A typical structure is: abstract; keywords; introduction; related work; methods; results and discussion; conclusion; Declaration on Generative AI; references. Position and interdisciplinary papers may depart from it where the argument requires.

–    Situate the work in the current international literature. A regular paper typically cites 25–30 sources, most of them from the last five to eight years, drawn from venues beyond the authors’ own institution and country.

–    Keep self-citation proportionate: as a rule below 15% of the reference list, and in no case above 20%.

–    Cite sources individually rather than in long undifferentiated strings; more than three consecutive citations for a single claim usually indicates that the claim should be stated more precisely.

–    Give a language tag for sources not published in English, for example “(in Ukrainian)”.

–    Format headings, captions, tables, and equations uniformly; supply figures as vector graphics wherever possible, with fonts that remain legible at print size.

General chair

Prof. Tetiana Hovorushchenko, Khmelnytskyi National University, Ukraine

Programme committee chairs

–   Prof. Oleksander Barmak, Khmelnytskyi National University, Ukraine

–   Prof. Iurii Krak, Taras Shevchenko National University of Kyiv, Ukraine

International programme committee

Additional reviewers with relevant expertise will be invited as needed.

Organising committee

–   Dr. Pavlo Radiuk, Khmelnytskyi National University, Ukraine –   Dr. Oleksandr Mazurets, Khmelnytskyi National University, Ukraine

–   Dr. Maryna Molchanova, Khmelnytskyi National University, Ukraine –   Dr. Olena Sobko, Khmelnytskyi National University, Ukraine

Proceedings editors

The CEUR-WS.org volume will be edited by Tetiana Hovorushchenko, Oleksander Barmak, Abdel-Badeeh M. Salem, and Pavlo Radiuk.

---

Acknowledgement. The Microsoft CMT service is used for managing the peer-review process for this workshop. This service is provided free of charge by Microsoft, which bears all expenses, including costs for Azure cloud services as well as for software development and support.

Explore the agenda of ExplAI-2025.

We invite original papers on (but not limited to) theoretical and practical advances in transparent, trustworthy, and ethical AI systems for decision-making aligned with the UN Sustainable Development Goals.

– Paper submission deadline: 20.09.2025
– Notification of acceptance: 20.10.2025
– Camera-ready submission: 27.10.2025
– Workshop date: 07.11.2025

– Paper type: Regular papers (10–15 pages), English.
– Template (CEUR-ART, one-column):
    – LaTeX ZIP: CEURART.zip (recommended with Overleaf)
    – Libre Office template: Download
– Authorship limits: ≤ 5 authors per paper.
– Self-citation: ≤ 20 % of references.
– Submission system: Submit via Microsoft CMT.
– Peer review: Double-blind. Each submission will receive at least two independent reviews. Acceptance requires at least one strong accept and no rejects; a third reviewer may be invited in case of disagreement.

Generative AI disclosure: Authors must include the mandatory CEUR declaration regarding the use (or non-use) of generative AI tools (see CEUR policy: https://ceur-ws.org/GenAI/Policy.html).

Author metadata: Meaningful affiliations with institutional e‑mail addresses and ORCID identifiers (https://orcid.org) must be provided in the paper footer (page 1), per CEUR-ART instructions.

To improve acceptance likelihood and ensure CEUR compliance:
– Concise, specific title (≤ 13 words; avoid repeated words).
– Clear English; please proofread.
– Paper structure: Abstract; Keywords; Introduction; Related Work; Methods/Model; Results & Discussion; Conclusion; References.
– Reference list: ~25–30 sources typical for regular papers; majority recent (≤ 5–8 years).
– International reach: ≤ 20 % total local (same-country + self) references; self-citation ideally ≤ 15 % and in all cases ≤ 20 %.
– No «cascade» chains of >3 sequential references.
– Non-English sources must include language tags (e.g., (in Ukrainian)).
– Ensure uniform formatting of headings, captions, and equations; provide high-quality figure graphics with readable fonts (vector graphics are preferred).

Each accepted paper must be accompanied by a signed CEUR-WS Open-Access Author Agreement (hand-completed and scanned/PDF):

AUTHOR‑AGREEMENT (NTP) – select this form if your submission contains no copyrighted third‑party material and no content generated with AI/LLM tools (the usual case).

AUTHOR‑AGREEMENT (TP) – select this form if your paper does include copyrighted third‑party material. Attach the corresponding permission letters to the signed agreement.

General Chair
– Prof. Tetiana Hovorushchenko, Khmelnytskyi National University, Ukraine

Program Committee Chairs
– Prof. Olexander Barmak, Khmelnytskyi National University, Ukraine
– Prof. Iurii Krak, Taras Shevchenko National University of Kyiv, Ukraine

International Program Committee
– Prof. Eduard Manziuk, Khmelnytskyi National University, Ukraine
– Prof. Nataliia Shakhovska, Lviv Polytechnic National University, Ukraine
– Prof. Vyacheslav Kharchenko, National Aerospace University KhAI, Ukraine
– Prof. Volodymyr Lytvynenko, Lublin University of Technology, Poland
– Prof. Vyacheslav Kovtun, Institute of Theoretical and Applied Informatics, Polish Academy of Sciences, Poland
– Prof. Roman Kuc, Yale University, USA
– Prof. Orken Mamyrbayev, Institute of Information and Computational Technologies, Kazakhstan
– Prof. Andrii Biloshchytskyi, Astana IT University, Kazakhstan
– Prof. Waldemar Wojcik, Lublin University of Technology, Poland
– Dr. Ivan Izonin, The Bartlett School of Sustainable Construction, University College London, UK
– Prof. Sergey Yakovlev, Lodz University of Technology, Poland
– Prof. Adam Wierzbicki, Polish-Japanese Academy of Information Technology, Poland
– Prof. Sergii Babichev, Jan Evangelista Purkyne University in Usti and Labem, Czech Republic
– Dr. Dmytro Chumachenko, University of Waterloo, Canada
– Prof. Vitaly Levashenko, Zilina University, Slovakia
– Prof. Elena Zaitseva, Zilina University, Slovakia
– Dr. Miroslav Kvassay, Zilina University, Slovakia
– Dr. Artem Boyarchuk, TalTech University, Estonia
– Dr. Tomáš Sochor, University of Ostrava, Czech Republic
– Prof. Abdel-Badeeh M. Salem, Ain Shams University, Egypt
– Dr. Houda El Bouhissi, University of Bejaia, Algeria

Additional reviewers with relevant expertise will be invited as needed.

Organizing Committee
– Dr. Pavlo Radiuk, Khmelnytskyi National University, Ukraine
– Dr. Oleksandr Mazurets, Khmelnytskyi National University, Ukraine
– Dr. Maryna Molchanova, Khmelnytskyi National University, Ukraine
– Dr. Olena Sobko, Khmelnytskyi National University, Ukraine

Authors of high-quality workshop papers will be invited to submit extended versions (≥ 70 % new content) to selected journals. Confirmed opportunity:

Sustainability, ISSN 2071–1050, Q2 (Scimago), IF 3.3. Special Issue: «Empowering Artificial Intelligence to Achieve Sustainable Development Goals». Guest Editors: Prof. Dr. Olexander Barmak, Prof. Dr. Iurii Krak, Prof. Dr. Eduard Manziuk, Dr. Pavlo Radiuk.

All extended submissions will undergo the journal’s standard peer-review process.

Questions? Please e-mail explai.khmnu@gmail.com

Updates: https://kn.khmnu.edu.ua/ExplAI-2025/