Generative artificial intelligence is commonly framed as a challenge to employment, authorship, academic integrity, professional competence and human intellectual distinctiveness. This paper argues that these apparently separate disputes share a deeper institutional problem. For much of modern educational and professional life, personal execution of an intellectual task provided a useful, though imperfect, signal of the capability underlying it. Generative AI weakens that inference because increasingly sophisticated cognitive execution can now be obtained externally. This paper describes institutional reliance on production as evidence of capability as execution-proxy dependence and develops a related empirical hypothesis, execution-command dissociation: as externally available AI performs more cognitive execution, the predictive relationship between the quality of an intellectual artefact and the producer's demonstrated command of that artefact may weaken. The paper develops substantive intellectual command as a more direct object of evaluation, distinguishes capability amplification from competence amplification, and examines implications for assessment, authorship, disclosure, professional competence and human agency. Its central proposition is that AI is weakening the informational value of execution as evidence of human capability. The resulting empirical claim is deliberately falsifiable: if artefact quality remains equally informative about independently demonstrated intellectual command under AI-assisted and unaided production, the theory must be revised.
Onosoji O. Ononuga· Zenodo (CERN European Organi...· 0 citations
Generative artificial intelligence is commonly framed as a challenge to employment, authorship, academic integrity, professional competence and human intellectual distinctiveness. This paper argues that these apparently separate disputes share a deeper institutional problem. For much of modern educational and professional life, personal execution of an intellectual task provided a useful, though imperfect, signal of the capability underlying it. Generative AI weakens that inference because increasingly sophisticated cognitive execution can now be obtained externally. This paper describes institutional reliance on production as evidence of capability as execution-proxy dependence and develops a related empirical hypothesis, execution-command dissociation: as externally available AI performs more cognitive execution, the predictive relationship between the quality of an intellectual artefact and the producer's demonstrated command of that artefact may weaken. The paper develops substantive intellectual command as a more direct object of evaluation, distinguishes capability amplification from competence amplification, and examines implications for assessment, authorship, disclosure, professional competence and human agency. Its central proposition is that AI is weakening the informational value of execution as evidence of human capability. The resulting empirical claim is deliberately falsifiable: if artefact quality remains equally informative about independently demonstrated intellectual command under AI-assisted and unaided production, the theory must be revised.
Onosoji O. Ononuga· Zenodo (CERN European Organi...· 0 citations
Bedside use of artificial intelligence (AI) platforms is increasingly common. Busy clinicians may welcome these generative AI (Gen AI) tools, which have the potential to streamline many time-consuming tasks and aid in patient care. Trainees may find them useful to quickly evaluate complex medical information. Acceptance of bedside Gen AI tools by patients and their families, however, is less clear, and a definitive standard surrounding informed consent has yet to be established. Omission of certain types of information by Gen AI tools, including "small talk," which comprises an essential element of many pediatric clinical interactions, demands attention alongside the tendency of Gen AI tools to fabricate content. Medical students and resident physicians may be early adopters of Gen AI tools and may accept AI-generated statements at face value before having fully developed adequate knowledge and critical skills to independently evaluate their veracity. Despite these challenges, Gen AI tools in the clinical space are here to stay. Safe and effective incorporation of these tools in patient care and medical education must balance a wide range of considerations. In the following Ethics Rounds, stemming from a 2024 Pediatric Academic Society Bioethics Club Meeting session on the use of AI, the commentators draw on their diverse background as pediatric clinicians, bioethicists, and educators to explore these issues.
Katherine E. MacDuffie, Douglas S. Diekema, Jennifer C. Kett et al.· Pediatrics· 0 citations
The manual creation of personalized travel itineraries remains a labor-intensive process that requires travel agents to consolidate heterogeneous information from multiple sources, including natural language interactions, booking confirmations, screenshots, and reservation documents. Although recent advances in multimodal artificial intelligence have significantly improved language understanding and content generation, existing solutions typically address isolated tasks rather than providing an integrated workflow capable of automating the complete travel planning process. This paper presents the development and validation of an AI-powered conversational agent designed to automate itinerary generation and generative video synthesis through natural language processing (NLP) and multimodal data extraction. The system, evaluated at Technology Readiness Level 4 (TRL4), employs a Multi-Agent System (MAS) architecture, integrating specialized large language models (Gemini 2.5/2.0 Flash, GPT-4o-mini) with optical character recognition (OCR) and Latent Diffusion Models to interpret user requests, extract structured data from images and PDFs, and produce comprehensive travel packages formatted as professional PDF deliverables alongside AI-synthesized promotional videos. Validation in a controlled laboratory environment demonstrated over 90% intent recognition accuracy, successful data extraction from non-standard formats, and the automated production of both client-ready documents and coherent visual narratives from static itinerary data. The system achieved an average video generation latency of 4.2±0.8 minutes, while maintaining structural consistency in the generated PDF itineraries. The system represents a viable proof-of-concept for intelligent travel planning automation, with implications for enhancing operational efficiency in the tourism industry and reducing cognitive load on human agents. Future work will advance the prototype to TRL5 through integration with external booking APIs and real-user testing scenarios.
Pablo Vicente-Martínez, Carlos Ferrer-Baixauli, Emilio Soria-Olivas et al.· Applied Sciences· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Selective photothermolysis is the basic theoretical foundation of laser therapy, and now a highly used diagnostic and treatment method in dermatology is based on it. Although laser therapy has shown some good curative effects on various skin diseases, clinical application still faces many problems: the parameters are not standardized, the criteria for choosing them are inconsistent, the assessment of efficacy is mostly based on subjective visual checks, and dark-skinned people are more prone to adverse effects such as burns, pigment changes, etc. The general process and applications of artificial intelligence in laser dermatology will be introduced in this paper, such as pre-operative quantitative analysis of lesion images, intelligent adjustment of laser energy during surgery, and objective quantification of the effect of post-operative treatment. There are many serious problems in the actual operation that have not been solved yet; there is a lack of high-level clinical validation data, an opaque "black-box" mechanism for algorithms, data bias due to racial imbalance, a fragmented industrial supervision system, unreliable content hallucination from large language models, etc. The four problems that need to be solved in the new round of targeted countermeasures are large-scale multicenter cohort studies, the development of interpretable AI models, fair and standardised patient data privacy governance, and optimisation of human-machine collaborative clinical workflows. The aims of this paper are to offer theoretical support for the standardisation of clinical application of AI-assisted laser intervention and to help realise the goal of personalised and precise skincare.
Yanhong Song· Theoretical and Natural Scie...· 0 citations
As artificial intelligence is increasingly applied to Chinese composition education, writing assessment is shifting from single-score judgment to semantic understanding and ability diagnosis. Student essays not only present language expression outcomes, but also contain evidence of learning behavior, including topic understanding, structural organization, content development, logical coherence, and language use. To address this issue, this study uses the public Chinese Essay Dataset For Pre Training and selects Qwen2.5 7B Instruct as the core model to construct a method for semantic feature extraction and mastery state prediction. The model first extracts five types of features from essay content, essay type, and grade level, including topic understanding, structural completeness, content richness, logical coherence, and language expression. The original writing quality ratings are then mapped into three mastery states: low mastery, medium mastery, and high mastery. The experimental results show that, after semantic features are added, Accuracy increases from 0.684 to 0.731, Macro F1 increases from 0.672 to 0.725, QWK increases from 0.642 to 0.712, and RMSE decreases from 0.681 to 0.599. These results indicate that semantic features can improve the accuracy and ordinal consistency of writing mastery prediction. The proposed method provides a feasible pathway for intelligent diagnosis, precise feedback, and personalized learning support in Chinese composition education.
Minghong Liu· Applied and Computational En...· 0 citations
Artificial intelligence (AI) has become a significant force in the field of quantitative trading because it extends traditional rule-based systems to areas such as adaptive prediction, dynamic configuration and automated execution. At the same time, the widespread adoption of machine learning, reinforcement learning, and workflows based on large language models in financial practice has raised concerns about data overfitting, data vulnerability, herding effects, systemic risk, and governance failure. This paper provides a comprehensive review and conceptual synthesis of the application of AI in quantitative trading by integrating recent literature on machine learning, prediction and portfolio optimization in financial markets, strategy mining driven by large language models, quantitative crisis management, and AI-based risk management. This paper follows the structure of actual trading processes and explores how AI supports data collection, feature engineering, model development, portfolio construction, and execution. It further identifies four representative risk categories: model risk, data dependency and quality risk, market liquidity and volatility risk, and operational and cybersecurity risk. Based on these findings, this paper proposes a multi-layered governance framework that combines robust model engineering, investment-level controls, implementation safeguards, human oversight, and regulatory coordination. This paper argues that compared to replacing human judgment with fully autonomous algorithms, the future of quantitative trading relies more heavily on building auditable systems that combine data discipline, model validation, risk control, and human oversight.
Zichong Long· Applied and Computational En...· 0 citations
The rapid expansion of the low-altitude economy (LAE)demands highly reliable Unmanned Aerial Vehicle (UAV) systems. However, traditional UAV control engineering suffers from severe time latency inherent in its serial architecture and is significantly amplified in complex urban environments. To address this challenge, 6G-enabled Integrated Sensing, Communication, and Computing (ISCC) architectures are being utilised to reduce transmission delays. Furthermore, Large Language Models (LLMs) are introduced to shift the design paradigm of control engineering, aiming to improve decision timeliness through predictive state estimation and reasoning. This article reviews the advantages of coupling 6G ISCC with LLMs. The synthesis indicates that this integration effectively minimises system latency by substituting iterative mathematical calculations with direct semantic prediction, its transition to real-world deployment faces severe engineering bottlenecks where the intrinsic "black-box" hallucinations of LLMs are paramount. This paper concludes that autonomous UAV operations necessitate a fundamental shift toward Trustworthy and Explainable AI (XAI), coupled with high-fidelity digital twin validation, to guarantee system stability in safety-critical domains.
Kaiyi Chen· Applied and Computational En...· 0 citations
Abstract In this chapter, we focus on a digital innovation which is not confined to a single application such as digital platforms, but on one – Artificial Intelligence or AI – which is in course of pervading the whole of the economy and society, and fully deserves the title of one (or several) general purpose technologies or GPTs. Estimates of its likely economic effects, particularly on productivity and labour demand, remain extremely varied and deeply uncertain. The risks which it imposes on individual groups, or even the whole human race, are still speculative. The European Union passed in 2024 an AI Act which seeks to guard against some of those risks. More prosaically, the possibility of monopolization of the supply of AI, for example by dominance of key inputs such as chips, ‘compute’ or Large Language Models, has receded as rivalry has extended to massive AI investments by several major platform companies, as well as the emergence of inventive western rivals and serious Chinese competitors. Finally, AI pervades even the field of regulation. There will not only be regulation of AI, but also regulation being done with AI, leading to a possible arms race between the AI tools deployed by regulatees and those available to regulators.
Robert Baldwin, Martin Cave, Martín Lodge· 0 citations
Artificial intelligence is usually analysed as a productivity technology, a labour-market shock, or a governance problem. This article argues that its deeper political significance may arise from a fourth channel: ontological displacement. Ontological displacement occurs when a technology weakens the socially recognised characteristics through which people understand human worth, competence, status, agency, and legitimate authority. Advanced AI can produce such displacement even before mass unemployment occurs, because credible machine substitutes may devalue expertise, educational scarcity, cognitive achievement, and the belief that productive contribution is a primary basis of social membership. The article develops an AI Ideological Conflict Framework that links capability shocks to perceived replaceability, distributional stress, meaning loss, legitimacy gaps, ideological mobilisation, and possible intrastate or interstate conflict. The framework integrates labour economics, meaningful-work research, political behaviour, sacred-values scholarship, AI governance, and futures studies. Using evidence-weighted horizon scanning, causal layered analysis, and a two-axis scenario matrix, four structural futures are developed: Managed AI Capitalism, Humanist Social Dividend, Algorithmic Oligarchy, and Cybernetic Commonwealth. The analysis argues that the most dangerous period is not necessarily a mature post-work society, but a transitional regime in which employment remains culturally necessary for dignity and income while human cognitive labour becomes less economically scarce. The paper forecasts that future ideological conflict is likely to centre on two questions: who owns the surplus generated by machine intelligence, and whether human beings should retain final authority when artificial systems outperform them. Policy implications include broadening status and social contribution beyond employment, dispersing AI rents, preserving meaningful human agency, strengthening transition institutions, and treating human sovereignty as a constitutional rather than merely technical design question.
Kwan Hong TAN· Zenodo (CERN European Organi...· 0 citations
This thesis evaluates the legal and regulatory framework governing FinTech firms' deployment of algorithmic credit scoring under the Nigeria Data Protection Act (NDPA) 2023. It critically examines data sovereignty, algorithmic bias, transparency, accountability, and the consent loopholes present in Section 37 of the NDPA. The research investigates institutional fragmentation across regulatory bodies, specifically the Nigeria Data Protection Commission (NDPC), Central Bank of Nigeria (CBN), and Federal Competition and Consumer Protection Commission (FCCPC). It provides a comparative analysis contrasting Nigerian data governance frameworks with international standards, including the EU General Data Protection Regulation (GDPR) and the EU Artificial Intelligence Act.
EZEKIEL KELLY· Zenodo (CERN European Organi...· 0 citations
DLC-OD is an object-detection dataset containing 295 hand-sketched digital logic circuit images collected from 20 participants. The dataset includes 1,203 manually annotated instances belonging to seven logic-gate classes: AND, NAND, NOR, NOT, OR, XNOR, and XOR. The images and bounding-box annotations are provided in YOLO format. In the deposited version, the dataset is divided into 206 training images containing 831 instances, 60 validation images containing 255 instances, and 29 test images containing 117 instances. The dataset was developed to support research on hand-drawn logic-gate detection, symbolic object recognition, explainable artificial intelligence, class activation mapping, and educational engineering applications. Drawing variations among participants provide differences in symbol shape, stroke thickness, orientation, and circuit layout.
Noha ElMasry, Fahima A. Maghraby, Mohamed Waleed Fakhr· Zenodo (CERN European Organi...· 0 citations
