Standards

Deepfake technology has moved well beyond research laboratories, with synthetic video and audio now being deployed to commit corporate fraud, impersonate executives, and manipulate employees at scale. Most detection systems in the research literature fail in three recurring ways: they perform well in controlled tests but poorly in real deployment; they handle only one modality at a time; and they require computing resources that most organisations do not possess. This research addresses these problems through a detection framework built on a fine-tuned EfficientNetV2-S visual backbone combined with three independent Radial Basis Function (RBF) Support Vector Machine (SVM) classifiers, covering visual, audio, and combined audio–visual detection, with the classification stage adding 2.3 ms per frame to a pipeline that requires 195 ms per frame on a single CPU thread. Evaluation followed the official FaceForensics++ source-video split, with all partitioning performed at video level. On FaceForensics++ C23, the visual classifier achieved an AUC-ROC (Area Under the Receiver Operating Characteristic Curve) of 0.9848 (95% CI 0.9787 to 0.9898) at the frame level and 0.9937 at the video level, with an accuracy of 95.53%. On a held-out DeepFakeDetection subset comprising unseen-manipulation methods and subjects, AUC-ROC was 0.9689. Cross-dataset evaluation on Celeb-DF v2 without retraining produced an AUC-ROC of 0.6589, a drop of 32.59 percentage points, with 65.6% of real frames flagged as fake at the inherited operating threshold; a recalibration experiment shows that no operating point on that curve is usable. On the full ASVspoof 2019 Logical Access evaluation partition, which contains thirteen attack types absent from training, the audio classifier reached an equal error rate of 16.55% (95% CI 14.70 to 18.40), ranging from 0.36% on the most detectable attack to 33.43% on the least. A controlled ablation on FakeAVCeleb v1.2, using identity-disjoint partitions and separate labels for each modality, showed a visual-only AUC-ROC of 1.0000 and audio-only of 0.9884. Score-level fusion reduced AUC-ROC to 0.9776 (DeLong p = 0.016), because averaging discards information when video and audio are manipulated independently. A governance framework is presented that maps each technical limitation to a specific organisational control, grounded in proportionality, independence, and adaptability and aligned to ISO/IEC 27001:2022 Annex A, the National Institute of Standards and Technology (NIST) Cybersecurity Framework 2.0 and the NIST AI Risk Management Framework. The detection results establish that this governance layer cannot depend on detector output once the deployment domain differs from the training domain.

Children exposed to violence require a comprehensive, multidisciplinary support network operating within a well-coordinated child protection system. The effectiveness of such systems depends fundamentally on strong intersectoral collaboration and high-quality relationships among professionals from different sectors. Enhancing system performance therefore requires systematic identification and mitigation of management barriers that hinder coordinated action. This study proposes a Multi-Criteria Decision Analysis (MCDA) model to identify, analyze, and prioritize management barriers within child protection networks through an integrated MCDA framework for complex systems. The Fuzzy Delphi Method (FDM) was used to screen the most relevant barriers, swing weighting was used to elicit the relative importance of the retained barriers, the Analytic Hierarchy Process (AHP), applied in its absolute measurement mode, was used to derive priorities for the predefined performance levels and support the sector-specific evaluation of three candidate decision alternatives, and Social Network Analysis (SNA) degree centrality was used to weight the judgments of the four participating experts (one per network sector) in the final aggregation. The proposed approach was implemented in the Child Protection Network of a medium-sized Brazilian municipality. The results show that the Unified Information System alternative achieved the highest overall priority (w = 0.3525), closely followed by the Creation of an Intersectoral Management Group (0.3450) and Continuous Training for Network Professionals (0.3025); given the narrow margin (approximately 0.0075) between the two top-ranked alternatives, this prioritization should be interpreted as an indicative rather than a definitive ranking. More broadly, the results demonstrate that (i) the proposed framework was successfully applied, providing a structured and transparent process for identifying and prioritizing critical management barriers, and (ii) the resulting prioritization supports decision-makers and practitioners in allocating resources more effectively, strengthening intersectoral coordination, and improving services for children and adolescents exposed to violence.

A redefinition of genetically modified organisms (GMOs) is presented by means of bioinformatics and artificial intelligence. Grounded on existing information, an AI framework for GMO definition and detection is proposed with quantifiable information entropy parameters, similarity metrics to feed deep learning AI systems. The proposed AI framework aims to offer a data reform paradigm to upgrade GMO definition for versatile safety assessments of current and emerging artificial organisms. It is envisioned to include recent biotechnological advances such as CRISPR editing, epigenetic modifications, and even mRNA technologies. As global GMO applications continue to expand and molecular genetic technologies evolve rapidly, updated definitions and regulatory policies upgraded in AI environments are urgently required.