EDITORIAL: Trustworthy Intelligence: Reliability, Transparency, and Responsibility in Soft Computing

Abstract

Advances in artificial intelligence and soft computing are frequently assessed through improvements in predictive accuracy, yet numerical performance alone provides an incomplete basis for judging scientific quality and practical value. This editorial examines the broader responsibilities involved in developing, evaluating, and publishing intelligent systems. Drawing on established literature and contributions featured in this issue of the International Journal of Advances in Soft Computing and its Applications, it discusses methodological rigor, reproducibility, generalization, explainability, fairness, and computational efficiency as complementary dimensions of trustworthy research. Attention is given to the justification of hybrid architecture, the importance of fair experimental comparisons, and the need to communicate uncertainty and application limitations. The editorial also addresses generative AI assistance in research and manuscript preparation, emphasizing transparent disclosure and continuing human accountability. Similarity and AI-writing reports are considered supporting instruments whose outputs require contextual interpretation rather than automatic judgments about originality or misconduct. Finally, the editorial considers the role of journals in strengthening research practice through constructive peer review, consistent standards, and interdisciplinary dialogue, while introducing a continuing series on soft computing applications. It argues that sustainable scientific progress depends on contributions that deepen understanding and demonstrate reliable, proportionate benefits under clearly stated conditions.