Open AccessPeer ReviewedDouble-Anonymous ReviewCC BY 4.0 ↗Continuous Publication
1.0 Aims and Scope
The Journal of AI, Automation & Intelligent Systems (JAAIS) publishes rigorous theoretical, computational, experimental and applied scholarship concerned with the development, evaluation and responsible deployment of artificial intelligence, automation and intelligent systems. The journal provides a forum for research that advances the scientific foundations of intelligent technologies while also examining how those technologies behave when transferred from controlled experimental environments into complex operational settings.
JAAIS is particularly interested in research that connects algorithmic innovation with system-level performance. Contributions should therefore move beyond demonstrations of isolated predictive accuracy and explain how intelligent systems perform under realistic constraints including incomplete or changing data, limited computational resources, uncertainty, latency, sensor degradation, human interaction, adversarial conditions, safety requirements and distribution shift.
The journal welcomes work spanning artificial intelligence, machine learning, deep learning, reinforcement learning, computer vision, intelligent control, robotics, autonomous systems, embodied intelligence, human–robot interaction, edge intelligence, Internet of Things systems, cyber-physical systems, digital twins and industrial automation. Research addressing intersections among these fields is particularly encouraged where the integration creates a substantive scientific or engineering contribution.
JAAIS also provides a venue for research on trustworthy intelligent systems. Relevant work includes robustness, reliability, explainability, uncertainty estimation, verification, validation, privacy-preserving learning, adversarial resilience, human oversight, algorithmic accountability, model monitoring and safety-critical AI.
The journal recognises that the practical significance of intelligent systems increasingly depends on more than model performance. It therefore welcomes research addressing the complete lifecycle of intelligent technologies, including data acquisition, system design, training, validation, deployment, monitoring, maintenance and adaptation.
Research may be theoretical, simulation-based, laboratory-based, field-deployed or mixed in design. However, claims must remain proportional to the evidence presented. Results obtained exclusively from simulation or benchmark datasets should not be described as evidence of successful real-world deployment unless such deployment has actually been evaluated.
1.1 Core Research Areas
The journal's core areas include machine learning and deep learning; reinforcement and adaptive learning; computer vision and multimodal perception; robotics and autonomous machines; human–robot and human–AI interaction; intelligent control; industrial automation; edge and embedded intelligence; IoT-enabled intelligence; cyber-physical systems; digital twins; sensor fusion; predictive and condition-based maintenance; autonomous decision systems; trustworthy and responsible AI; explainable AI; model robustness; uncertainty quantification; AI safety; privacy-preserving intelligent systems; distributed intelligence; intelligent manufacturing; and real-world evaluation of AI-enabled technologies.
JAAIS also considers emerging research that falls between conventional disciplinary boundaries where the manuscript demonstrates clear relevance to intelligent systems.
1.2 Interdisciplinary and Emerging Areas
Interdisciplinary work may connect artificial intelligence with engineering, manufacturing, healthcare technology, environmental monitoring, mobility, agriculture, infrastructure, energy systems, public services or other application domains. Such articles must nevertheless make a genuine contribution to AI, automation or intelligent-systems scholarship rather than merely applying an existing algorithm to a new dataset.
The journal particularly welcomes research investigating how intelligent systems interact with people, institutions, physical infrastructure and changing environments.
2.0 Contributions We Welcome
JAAIS considers Original Research, Review Articles, Systematic Reviews where appropriate, Methods and Technical Articles, Short Communications, Perspectives, Case Studies, Editorials and other article types approved by the journal.
Original Research should present a clearly defined research problem, appropriate methodology, sufficient comparison with relevant baselines and conclusions supported by the reported evidence.
Review articles should provide analytical synthesis rather than descriptive catalogues of prior studies. Methods and Technical Articles should demonstrate why the proposed technique matters and provide sufficient detail for independent evaluation or replication.
2.1 Evidence and Methodological Expectations
Authors should report datasets, sampling or data-generation procedures, preprocessing, training and validation strategies, evaluation metrics, baselines, statistical procedures and computational resources sufficiently clearly to permit expert assessment.
Claims of superior performance should be supported by appropriate comparisons. Where stochastic models are used, authors should report variability rather than a single favourable run where technically feasible.
Authors should distinguish training, validation and test data clearly and address potential leakage or inappropriate reuse of benchmark information.
3.0 Disciplinary Standards
Reproducibility is a central expectation of JAAIS. Authors should provide sufficient information concerning algorithms, hyperparameters, software environments, hardware, model architecture, experimental conditions and evaluation procedures.
Where source code, trained models or datasets can legitimately be shared, authors are encouraged to provide stable repository links. Where they cannot be shared, the limitation should be explained.
Research involving physical robots, autonomous platforms or cyber-physical systems should describe hardware configuration, sensing, control architecture, experimental environment and safety considerations.
3.1 Research Transparency
Authors should distinguish clearly between measured results, simulated results, inferred results and projected performance. Computational cost, latency, memory requirements and energy consumption should be reported where they materially affect claims about practical deployment.
Bias, privacy, cybersecurity, robustness and safety implications should be addressed when relevant to the system being studied.
4.0 Topics Outside Scope
JAAIS does not normally consider manuscripts consisting primarily of routine application of an established algorithm to a new dataset without methodological or substantive insight; promotional descriptions of commercial products; unsupported claims about artificial intelligence capabilities; manuscripts lacking sufficient methodological information for meaningful evaluation; or speculative commentary presented as empirical evidence.
From this journal
JAAIS editorials
These four editorials develop the journal’s research priorities and the questions that define its field.
Intelligent systems are moving rapidly from controlled experimental environments into operational settings where their decisions influence people, infrastructure, machines, and physical processes. Yet strong performance on curated benchmarks provides only limited evidence that such systems will remain dependable when confronted with distribution shift, incomplete data, sensor degradation, adversarial manipulation, human interaction, or changing operational conditions. Trustworthy deployment therefore requires a broader evidentiary standard than predictive accuracy alone. This editorial argues that operational trustworthiness is an emergent system-level property produced through the interaction of robustness, uncertainty awareness, safety engineering, explainability, cybersecurity, human oversight, governance, and lifecycle monitoring. Particular attention is given to the benchmark-to-deployment gap, failure-condition validation, inappropriate human reliance, adversarial vulnerabilities, and accountability across complex development and supply chains. The field should prioritize realistic stress testing, evaluation under distribution shift, uncertainty-aware decision architectures, transparent deployment reporting, continuous monitoring, human-in-the-loop safeguards, incident learning, and interoperable safety standards. Progress should ultimately be judged by whether intelligent systems remain reliable, understandable, and controllable under the conditions in which society actually depends on them.
Autonomous and intelligent systems increasingly work alongside people in factories, vehicles, healthcare environments, service settings, and decision-support contexts. Their success, however, should not be defined by the extent to which human involvement can be eliminated. Highly autonomous systems may perform individual functions accurately while still producing poor operational outcomes when users misunderstand system capabilities, lose situational awareness, become over-reliant, cannot intervene effectively, or experience excessive cognitive workload. This editorial argues that successful autonomy should be evaluated through the quality of human–machine collaboration rather than maximum machine independence. Human-centred autonomy requires coordinated attention to trust calibration, shared control, situation awareness, cognitive ergonomics, explainability, communication, safe transfer of authority, inclusive design, training, and organizational responsibility. Current evidence also demonstrates that explanations, anthropomorphic interfaces, and greater automation do not automatically improve collaboration. Research should therefore prioritize standardized human-autonomy evaluation, longitudinal trust studies, adaptive and multimodal interfaces, safe control transitions, representative end-user testing, and interdisciplinary validation. Autonomous systems should ultimately augment human capability while preserving meaningful understanding, agency, intervention, and responsibility.
Intelligent automation increasingly depends on computation occurring close to sensors, machines, and physical processes rather than exclusively in centralized cloud infrastructures. Industrial Internet of Things deployments, embedded artificial intelligence, autonomous devices, digital twins, and cyber-physical systems generate continuous data streams whose operational value often depends on low latency, reliable connectivity, privacy preservation, energy efficiency, and immediate local response. This editorial argues that edge intelligence should therefore be understood as a fundamental architectural component of next-generation automation rather than merely a reduced form of cloud computing. The emerging landscape combines distributed inference, embedded machine learning, federated learning, real-time networking, digital twins, predictive maintenance, and edge-cloud coordination, but substantial challenges remain in constrained computation, heterogeneous hardware, interoperability, cybersecurity, synchronization, model lifecycle management, fault tolerance, and distributed-system validation. Future progress requires energy-efficient models, hardware-software co-design, resilient communication and control, secure edge architectures, standardized benchmarking, trustworthy federated learning, and validation under realistic hardware and network conditions. Dependable automation will ultimately require coordinated design across sensing, communication, computation, intelligence, and physical control.
Adrian Mercer, Leila Rahman, Daniel Okafor, Sofia Petrov
Rapid advances in artificial intelligence and robotics are scientifically meaningful only when reported results can be independently understood, evaluated, compared, and, where feasible, reproduced. Yet contemporary research remains vulnerable to undocumented preprocessing, inaccessible datasets, hidden hyperparameter choices, random-seed sensitivity, weak baselines, benchmark overfitting, selective reporting, incomplete statistical analysis, simulation-to-reality gaps, and insufficient disclosure of computational resources. This editorial argues that reproducibility should be treated as a foundational condition of cumulative progress rather than an optional feature of exemplary research. Reproducibility, replicability, repeatability, robustness, and generalizability describe related but distinct dimensions of scientific reliability, each requiring appropriate evidence. The field should move toward transparent experimental reporting, standardized benchmark protocols, multiple-run evaluation, uncertainty quantification, stronger baseline comparisons, systematic ablation, accessible code and data where feasible, computational-resource disclosure, external validation, and real-world robotic testing. Journals, conferences, reviewers, and research institutions also shape these practices through incentives and publication standards. Scientific progress in intelligent systems should ultimately be measured not only by higher benchmark scores, but by whether claims remain credible when independently examined.
