Call for papers: Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems (Engineering Applications of Artificial Intelligence, Elsevier)
A special issue of Engineering Applications of Artificial Intelligence invites work on AI that detects and resists cyber threats while keeping people meaningfully in the loop, with submissions open from 1 October 2026 to 31 January 2027 (Elsevier, 2026). Here is what the guest editors ask for, how to judge fit and how to plan a credible paper in under four months.
PRPhD Research Lab InsightsEditorial team, reviewed by a subject specialist
Engineering Applications of Artificial Intelligence (ISSN 0952-1976; online 1873-6769)
Publisher
Elsevier
Special issue
Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems
Peer review
Double anonymised; editor screening, then typically at least two reviewers
Journal article types
Original research articles, short communications, survey papers or tutorials, case studies or software reviews
Access model
Hybrid: subscription route with no publication fee, or open access with an article publishing charge
APC
USD 3,040 excluding taxes, as stated on the journal homepage on 5 October 2026
Publication
Accepted articles go online immediately, appear in a regular issue and are collected in the online special issue
Guest editors
Dr. Stefano Cirillo (Leading Guest Editor)Università degli Studi di Salerno, Salerno, Italy
Dr. Idio GuarinoAlma Mater Studiorum Università di Bologna, Bologna, Italy
Dr. Pantaleone NespoliUniversidad de Murcia, Murcia, Spain
Dr. Eliana PastorPolitecnico di Torino, Turin, Italy
Dr. Giancarlo SperlìUniversità degli Studi di Napoli Federico II, Naples, Italy
Dr. Serena TardelliConsiglio Nazionale delle Ricerche, Rome, Italy
Topics the issue invites
Natural language processing and LLM-based methodology for identifying malicious, deceptive and manipulative content
Data-driven and knowledge discovery approaches for cyber threat detection in large-scale heterogeneous environments
Graph learning and knowledge graph techniques for modelling cyber threat ecosystems, communication networks and social interactions
Generative AI for both offensive and defensive cybersecurity purposes
AI-driven cyber threat intelligence, attack attribution and predictive threat analytics
Human-AI interaction, trust calibration, automation bias, overreliance and decision support systems
Human-centred approaches to combating phishing, social engineering and disinformation
Human-in-the-loop and hybrid intelligence systems for dynamic risk assessment and explainability
Behavioural, cognitive and organisational factors influencing cybersecurity awareness and resilience
Real-world deployments of AI-based solutions
Industrial case studies, empirical evaluations and benchmarking of AI-driven security approaches
Cross-domain applications in critical infrastructures, healthcare, smart cities, transportation and Industry 5.0
Lessons learned, best practices and evidence-based assessments of trustworthy AI systems in real settings
Is your work a fit? Check before you write
Your paper makes an AI contribution AND shows an engineering application; the journal asks the abstract to state both explicitly (Elsevier, n.d.-c).
The security problem sits in a cyber-physical or socio-technical setting (critical infrastructure, healthcare, transport, smart cities, industry), not only on a generic benchmark (Elsevier, 2026).
Your work addresses at least one of the two levels the call names: model-level attacks on AI, or human-centred threats such as phishing and disinformation (Elsevier, 2026).
You can state a realistic threat model and evaluate against it, rather than reporting accuracy alone (Apruzzese et al., 2023).
You can deposit data and code, or explain why not, under the journal's data policy (Elsevier, n.d.-c).
You can submit a complete, single-column manuscript by 31 January 2027 with all co-authors' agreement.
Journal metrics
9.0Impact FactorAs displayed on the journal homepage, checked 5 Oct 2026; the homepage does not print the JCR year next to the figure (Elsevier, n.d.-a)
11.7CiteScoreAs displayed on the journal homepage, checked 5 Oct 2026 (Elsevier, n.d.-a)
IndexedScopusListed under indexing on the journal's ScienceDirect insights page; Scopus-based SCImago record shows coverage 1988 to 2026 (Elsevier, n.d.-b; SCImago, n.d.)
1.782SJR 2025SCImago Journal & Country Rank (Scopus data), checked 5 Oct 2026 (SCImago, n.d.)
27 daysSpeedJournal-wide average from submission to first decision; 85 days to decision after review; 207 days to acceptance (Elsevier, n.d.-a)
Why this call is a priority now
AI is now part of everyday cyber defence. A systematic review of 236 primary studies mapped AI use cases across the NIST cybersecurity framework and called for work on emerging application areas, data representation and the infrastructure needed to adopt AI-based security in practice (Kaur et al., 2023). At the same time, the AI itself has become a target: the NIST taxonomy of adversarial machine learning catalogues poisoning, evasion, privacy and misuse attacks across the AI lifecycle, including attacks on generative models (Vassilev et al., 2025).
The difficulty is that research and practice have drifted apart. Apruzzese et al. (2023) analysed recent adversarial ML papers in top security venues alongside real-world case studies and found that actual attackers often use simple, cheap tactics, while much research assumes powerful gradient-based adversaries; they argue for precise, cost-driven threat models and reproducible evaluation. A survey of more than 300 papers on explainable AI in cybersecurity similarly warns that entrusting crucial security decisions to systems that cannot explain themselves is risky, and identifies open challenges in how explanations are produced and evaluated (Capuano et al., 2022).
The human side is the other gap. A systematic review of human factors in phishing shows that behavioural and cognitive factors remain central to why attacks succeed (Desolda et al., 2022). Regulation is raising the stakes: the EU AI Act requires high-risk AI systems to achieve appropriate accuracy, robustness and cybersecurity and to be resilient against attempts to manipulate them (Regulation (EU) 2024/1689), and the NIS 2 Directive sets a common level of cybersecurity for essential and important entities across the Union (Directive (EU) 2022/2555). Work that joins robust models, calibrated human trust and real deployments is what this call asks for.
Who should submit
PhD specialisations
Machine learning and adversarial robustness
Cybersecurity, intrusion and anomaly detection
Natural language processing for phishing, fraud and disinformation detection
Human-computer interaction and human-centred AI
Industrial control, smart grid and cyber-physical systems engineering
Graph learning and knowledge graphs for threat intelligence
Industry and sector fit
Energy, water and other critical infrastructure operators
Healthcare and medical device security
Transport, automotive and smart mobility
Manufacturing and Industry 5.0
Security operations centres and managed security providers
Countries and regions
European Union member states, where the AI Act sets robustness and cybersecurity duties for high-risk AI (Regulation (EU) 2024/1689) and NIS 2 sets cybersecurity measures for essential entities (Directive (EU) 2022/2555).
Italy and Spain, where all six guest editors are based, so the editors will know regional infrastructure and datasets well (Elsevier, 2026).
United States, where NIST's adversarial ML taxonomy offers shared terminology that reviewers will recognise (Vassilev et al., 2025).
Any country with a growing industrial IoT or smart city base, where real deployment data and case studies are the scarce resource the call values.
What novelty the editors expect
Threat models that reflect how real attackers behave, with costs and capabilities stated, rather than only white-box gradient attacks (Apruzzese et al., 2023).
Evaluation of explanations with the people who use them, such as analysts or operators, instead of reporting explanation methods without user evidence (Capuano et al., 2022).
Measurement of trust calibration, automation bias or overreliance in a security decision task, linked to an AI model's actual error profile (Elsevier, 2026).
Detection of AI-generated phishing or disinformation that is tested on new, out-of-distribution campaigns, not only on the dataset used for training.
Deployment or field studies in a cyber-physical setting, with lessons learned, which reviews identify as harder to find than laboratory results (Kaur et al., 2023).
Joint treatment of model-level and human-level risk in one system, the integration the call states as its aim (Elsevier, 2026).
Paper ideas you could develop
Robustness under realistic attackers
Take a public industrial control or IoT intrusion dataset, define a cost-bounded attacker who can only manipulate observable features, and compare detector robustness with and without adversarial training. Report the threat model in NIST terminology (Vassilev et al., 2025).
Explanations analysts actually use
Build a SHAP or counterfactual explanation layer for an alert triage model, then run a small within-subjects study with security analysts measuring decision accuracy, time and overreliance.
LLM-generated phishing in critical sectors
Assemble a corpus of human-written and LLM-generated phishing aimed at a sector such as healthcare, train detectors on one generator and test on another, and analyse which cues humans and models miss (Desolda et al., 2022).
Graph-based threat intelligence
Model indicators, campaigns and infrastructure as a knowledge graph, predict links to likely next targets, and evaluate attribution quality with temporal splits so the model never sees the future.
Your plan to the deadline
By 16 Oct 2026
Read the call and guide for authors; write the abstract stating the AI contribution and the engineering application in separate sentences (Elsevier, n.d.-c).
By 31 Oct 2026
Fix the threat model, datasets, baselines and evaluation metrics; secure ethics approval now if a user study is planned.
By 30 Nov 2026
Run the main experiments and robustness tests; log seeds, versions and settings for reproducibility.
By 15 Dec 2026
Complete ablations, any user evaluation and error analysis; freeze results.
By 4 Jan 2027
Draft the full manuscript in single-column format, with no undefined acronyms in the title or abstract (Elsevier, n.d.-c).
By 15 Jan 2027
Internal review by co-authors and a colleague outside the project; deposit data and code; prepare the generative AI declaration.
By 24 Jan 2027
Submit, selecting the article type "AI for Cybersecurity" (Elsevier, 2026), leaving a week before the 31 January deadline for system or co-author delays.
What you gain from publishing here
Publication in a Scopus-indexed journal with a long publication record, which supports visibility in the AI and engineering communities (SCImago, n.d.).
Double anonymised review by at least two reviewers, which gives structured expert feedback even if the paper is not accepted (Elsevier, n.d.-c).
A peer-reviewed article that can form a thesis chapter or support a thesis by publication, subject to your university's rules.
Placement in a themed collection that readers working on trustworthy AI for security are likely to browse (Elsevier, 2026).
A choice between a no-fee subscription route and paid open access, so funding need not decide whether you submit (Elsevier, n.d.-a).
What happens after you submit
The editors first assess whether the submission is suitable; unsuitable or out-of-format papers can be returned without review (Elsevier, n.d.-c).
Suitable papers typically go to at least two independent reviewers under double anonymised review (Elsevier, n.d.-c).
You receive a decision; most papers that progress need at least one round of revision with a point-by-point response.
If you disagree with a rejection, the journal allows one formal appeal per submission under Elsevier's policy (Elsevier, n.d.-c).
Once accepted, the article is posted online immediately, published in a regular issue within weeks and collected in the special issue (Elsevier, 2026).
You choose the subscription route or open access with the article publishing charge, checking any institutional agreement first (Elsevier, n.d.-a).
What the issue wants
The call starts from a simple observation: AI is moving into everyday applications and critical infrastructure, and this brings systemic risks at both the model and the socio-technical level. At the model level, it names adversarial examples, data poisoning, backdoor attacks, model extraction and membership inference. At the socio-technical level, it points to adversaries who target people through social engineering and disinformation, exploiting cognitive biases to compromise human-in-the-loop systems (Elsevier, 2026).
The guest editors want papers at the meeting point of AI, cybersecurity, human factors and risk management, from machine learning, knowledge discovery and human-centred computing. The topics fall into three groups: AI models for detecting, preventing and mitigating threats; human-centred AI for cybersecurity; and engineering applications and case studies in real settings (Elsevier, 2026). A paper that touches only one group can fit, but the strongest submissions will connect a technical contribution with evidence about people or deployment.
How to judge fit with this journal
Engineering Applications of Artificial Intelligence is an applied journal. Its guide for authors asks the abstract to state clearly which part is the contribution in AI and which is the application in engineering, and requires a single-column format, with desk rejection for papers that do not follow it (Elsevier, n.d.-c). A purely theoretical robustness bound, or a benchmark study with no engineering setting, is therefore a weaker fit than the same idea tested on a power grid, a hospital network or a vehicle system.
Ask, too, whether your evaluation would convince a practitioner. Apruzzese et al. (2023) found that many adversarial ML papers assume attackers with capabilities that real adversaries rarely use, and recommend stating threat models precisely and in terms of cost. Using shared vocabulary from the NIST taxonomy makes your assumptions easy for reviewers to check (Vassilev et al., 2025).
If your work is mainly about explanation, show that the explanations help someone make a better decision. Surveys of explainable AI in cybersecurity note that many methods are proposed without strong evidence of how they are evaluated or used (Capuano et al., 2022), and the call explicitly lists trust calibration, automation bias and overreliance as topics (Elsevier, 2026).
Building a credible evaluation in the time left
From 5 October 2026 to the 31 January 2027 deadline there are just under four months. That is realistic for a paper built on data and code you already have, but tight for new data collection or a user study that still needs ethics approval. Decide in the first fortnight whether your project is ready; if it is not, a later call may serve you better.
Use temporal or out-of-distribution splits so that results reflect how threats change over time.
Report robustness against a defined attacker, not just clean accuracy.
Compare against strong, recent baselines and explain why they were chosen.
If people are involved, measure their decisions, not only their stated trust.
Deposit data and code, or explain clearly why they cannot be shared, as the journal's data policy expects (Elsevier, n.d.-c).
The journal also asks authors to declare any use of generative AI tools in preparing the manuscript (Elsevier, n.d.-c). Keep a short record as you work so the declaration is accurate and easy to write.
What to do now
Open the official call and the guide for authors, and record the date you checked them (Elsevier, 2026).
Write two sentences: one naming your AI contribution, one naming the engineering application; if either is weak, rethink the framing (Elsevier, n.d.-c).
Map your paper to one of the three topic groups and state your threat model in recognised terms (Vassilev et al., 2025).
Decide early between the subscription route and open access, and check whether your institution covers the charge (Elsevier, n.d.-a).
Set an internal submission date of 24 January 2027 to leave a buffer before the deadline.
Your next step
Ready to submit to this call?
1Check the fit list above
2Get a pre-submission review if you want one
3Submit on the journal system before 31 January 2027
Can I submit a survey or tutorial to this special issue?
The journal publishes survey papers and tutorials among its article types (Elsevier, n.d.-c), but the call stresses real deployments, empirical evaluations and lessons learned (Elsevier, 2026). A survey should offer a clear framework or evaluation agenda rather than a list of papers. If unsure, ask the editorial office before writing.
Do I have to pay to publish?
No. The journal is hybrid: you can publish under the subscription model with no publication fee, or choose open access with an article publishing charge, stated as USD 3,040 excluding taxes on the journal homepage when we checked (Elsevier, n.d.-a).
Is a special issue paper reviewed less strictly?
It should not be. The call states that all manuscripts will be peer reviewed and judged on originality, significance, technical quality and clarity (Elsevier, 2026), and the journal uses double anonymised review (Elsevier, n.d.-c).
My work uses LLMs to detect phishing. Is that in scope?
Yes, if it sits in a cyber-physical or socio-technical setting. The call lists NLP and LLM-based methods for identifying malicious, deceptive and manipulative content, and human-centred approaches to phishing and social engineering (Elsevier, 2026).
Sources
Apruzzese, G., Anderson, H. S., Dambra, S., Freeman, D., Pierazzi, F., & Roundy, K. (2023). "Real attackers don't compute gradients": Bridging the gap between adversarial ML research and practice. In 2023 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML) (pp. 339–364). IEEE. https://doi.org/10.1109/SaTML54575.2023.00031
Capuano, N., Fenza, G., Loia, V., & Stanzione, C. (2022). Explainable artificial intelligence in cybersecurity: A survey. IEEE Access, 10, 93575–93600. https://doi.org/10.1109/ACCESS.2022.3204171
Desolda, G., Ferro, L. S., Marrella, A., Catarci, T., & Costabile, M. F. (2022). Human factors in phishing attacks: A systematic literature review. ACM Computing Surveys, 54(8), 1–35. https://doi.org/10.1145/3469886
Directive (EU) 2022/2555 of the European Parliament and of the Council of 14 December 2022 on measures for a high common level of cybersecurity across the Union (NIS 2 Directive). Official Journal of the European Union, L 333. https://eur-lex.europa.eu/eli/dir/2022/2555/2022-12-27/eng
Kaur, R., Gabrijelčič, D., & Klobučar, T. (2023). Artificial intelligence for cybersecurity: Literature review and future research directions. Information Fusion, 97, 101804. https://doi.org/10.1016/j.inffus.2023.101804
Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). Official Journal of the European Union, L, 2024/1689. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=OJ%3AL_202401689
Vassilev, A., Oprea, A., Fordyce, A., Anderson, H., Davies, X., & Hamin, M. (2025). Adversarial machine learning: A taxonomy and terminology of attacks and mitigations (NIST AI 100-2e2025). National Institute of Standards and Technology. https://doi.org/10.6028/NIST.AI.100-2e2025
How to cite this page
Choose your referencing style, then copy.
PhD Research Lab Insights. (2026, October 5). Call for papers: Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems (Engineering Applications of Artificial Intelligence, Elsevier). PhD Research Lab. https://www.phdassistance.com/insights/call-for-paper/trustworthy-ai-cybersecurity-cyber-physical-systems/
PhD Research Lab Insights. “Call for Papers: Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems (Engineering Applications of Artificial Intelligence, Elsevier).” PhD Research Lab, 5 Oct. 2026, www.phdassistance.com/insights/call-for-paper/trustworthy-ai-cybersecurity-cyber-physical-systems. Accessed [access date].
PhD Research Lab Insights (2026) Call for papers: Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems (Engineering Applications of Artificial Intelligence, Elsevier). PhD Research Lab. Available at: https://www.phdassistance.com/insights/call-for-paper/trustworthy-ai-cybersecurity-cyber-physical-systems/ (Accessed: [access date]).
PhD Research Lab Insights. “Call for Papers: Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems (Engineering Applications of Artificial Intelligence, Elsevier).” PhD Research Lab. October 5, 2026. https://www.phdassistance.com/insights/call-for-paper/trustworthy-ai-cybersecurity-cyber-physical-systems.
PhD Research Lab Insights, ‘Call for Papers: Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems (Engineering Applications of Artificial Intelligence, Elsevier)’ (PhD Research Lab, 5 October 2026) <https://www.phdassistance.com/insights/call-for-paper/trustworthy-ai-cybersecurity-cyber-physical-systems> accessed [access date].
PhD Research Lab Insights. Call for papers: Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems (Engineering Applications of Artificial Intelligence, Elsevier) [Internet]. PhD Research Lab; 2026 Oct 5 [cited [access date]]. Available from: https://www.phdassistance.com/insights/call-for-paper/trustworthy-ai-cybersecurity-cyber-physical-systems/
PhD Research Lab Insights, “Call for papers: Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems (Engineering Applications of Artificial Intelligence, Elsevier),” PhD Research Lab, Oct. 5, 2026. [Online]. Available: https://www.phdassistance.com/insights/call-for-paper/trustworthy-ai-cybersecurity-cyber-physical-systems. [Accessed: [access date]].
@online{trustworthy_ai_cybersecurity_cyber_physical_systems_2026, author = {{PhD Research Lab Insights}}, title = {{Call for papers: Human-Centered and Trustworthy AI for Cybersecurity in Cyber-Physical Systems (Engineering Applications of Artificial Intelligence, Elsevier)}}, organization = {PhD Research Lab}, date = {2026-10-05}, url = {https://www.phdassistance.com/insights/call-for-paper/trustworthy-ai-cybersecurity-cyber-physical-systems}, urldate = {[access date]} }