Development Of AI Research Articles

Generative AI in the Information Society: Implications for Higher Education and Research

<p dir="ltr"><span>Generative Artificial Intelligence (GenAI) represents a fundamental shift in AI development, moving from rule-based systems to neural networks capable of creating novel content and solving complex problems through pattern recognition and contextual understanding. This evolution challenges traditional Computer Science (CS) paradigms, as evidenced by innovations in large language models and diffusion-based image generation. This paper investigates how GenAI's emergence affects education and research in computer science and related fields. Through White's cultural model—examining technological, societal, and institutional dimensions—we analyse how GenAI's capabilities diverge from traditional CS approaches in both theory and practice. Our research reveals specific challenges for higher education, including the need to teach contextual reasoning, handle emergent behaviors, and develop adaptive problem-solving skills. We propose educational strategies such as project-based learning with GenAI tools and cross-disciplinary integration. These recommendations aim to establish GenAI as a distinct academic discipline while preparing students and researchers for its increasing role in scientific and professional practices.</span></p><div><span><br /></span></div>

Development of Large Language Models: Copyright Law Perspectives for Research Institutions and Research Libraries

Abstract This article discusses European copyright law as applied to the development and training of generative AI and natural language processing in public interest research institutions and libraries. The article focuses on the scope of the new exceptions from copyright law for text and data mining (TDM) for research purposes and discusses them from the perspective of research ethics and principles of open science in publicly financed research. The public interest mission of research institutions and libraries includes the open dissemination of research results but the exceptions from copyright are focused only on the training phase in AI development. Regulation on data transparency is fragmented. The article finds that while new exceptions open for developing language models under research institutions and libraries’ public interest mission to preserve national languages, the regulation is not adapted to principles of research ethics and open science, and legal uncertainty remains.

Machine Learning Utilization in Oncology for Precise Diagnosis and Informed Treatment

Machine learning (ML) technology has rapidly advanced and transformed sectors of today's world, healthcare specifically oncology, has benefited from such technology in recent years more than any other sector. This paper reviews the implementation of ML approaches in the field of oncology, specifically for precision diagnosis and personalized treatment, both of which are vital for dealing with the high level of complexity and heterogeneity associated with cancer. ML models are improving the accuracy of cancer diagnoses and prediction of patient-specific responses to treatments and enabling novel target discovery. Yet, the adoption of ML approaches in oncology is not devoid of difficulties. Such systems come with challenges such as data privacy, ethics, algorithmic biases, and technicalities such as model explainability and deployment in existing medical infrastructure, which this paper attempts to address However, the ML tools have a huge potential to change the way we practice and help us treat patients more efficiently, effectively and in a more patient-centered way. Then, the article discusses what the future holds for ML-driven oncology, such as big data capabilities, improvements in deep learning, and to for interdisciplinary collaborations between AI developers, oncologists and researchers. As this landscape continues to evolve, such considerations and innovation will be key in addressing the relevant ethical and practical challenges of machine learning to deliver on the promise of transforming cancer diagnosis and treatment.

Location is All You Need: Copyright Extraterritoriality and Where to Train Your AI

The development of artificial intelligence (“AI”) models requires vast quantities of data, which will often include copyrighted materials. The reproduction of copyrighted materials in the course of training AI models will infringe on copyright, unless there are applicable exceptions and limitations exempting such activities. There is so far considerable divergence between jurisdictions, including between the United States, EU, U.K., Japan, Singapore, Australia, India, Israel, and many more countries, in this regard. In the absence of international harmonization, there is therefore a high likelihood that the same type of training activity would be considered copyright infringement in some countries but not in others. The AI community is not blind to that risk. If copyright law restricts the development and deployment of AI, developers may decide to relocate their operations elsewhere, where the reproduction of training data is clearly not infringing. This Article concludes that there is a loophole in the international copyright system, as it currently stands, that would permit large-scale copying of training data in one country where this activity is not infringing. Once the training is done and the model is complete, developers could then make the model available to customers in other countries, even if the same training activities would have been infringing if they had occurred there. Because copyright laws are territorial in nature, by default they can only restrict infringing conduct occurring in their respective countries. From that point of view for AI developers, location is indeed all you need. The EU has become the first to respond to this problem by retroactively extending their text and data mining exception extraterritorially to training activities occurring in non-EU countries, once the completed AI model is placed on the EU market. While such an extraterritorial application benefits rightholders and closes the loophole now present, it makes the situation significantly more complex for developers. If other regulators decide to follow the same path as the EU, which previously happened in the data privacy context, then developers would be facing multiple, conflicting copyright laws targeting the same underlying activity. This could significantly complicate the development process for AI and potentially undermine the AI industry. This Article critically discusses these and related issues, and whether an extraterritorial application of copyright laws is compatible with territoriality norms that are supposed to respect foreign sovereignty. It also explores, in light of these difficulties, whether we should instead shift focus from regulating the inputs (i.e., the data used to train AI models) to regulating the outputs (i.e., the AI-generated content itself). Indeed, to the extent that the transnational data loophole cannot be closed without infringing upon foreign sovereignty, we may need to look at other regulatory means instead. The Article also suggests that we should consider model training and copyright infringement as a product-by-process problem, which calls for a comparison with how patent law solved similar extraterritoriality issues. Several decades ago, international patent treaties harmonized the extent to which patent laws can be applied extraterritorially to reach imported products derived from foreign manufacturing processes. If regulators wish to extend their copyright laws’ extraterritoriality to close the loophole that exists for training activities in the context of AI, and to do so in a way that is aligned with copyright territoriality, there may be a need to similarly revise international copyright treaties. This Article, therefore, urgently calls for a similarly coordinated international effort in copyright law, which balances the interests of rightholders with the technical, regulatory, and economic realities faced by developers. How we resolve these issues could make or break the future of AI. If we cannot find a way to reconcile the interests of rightholders and AI stakeholders, the world may be left with a segregated and fragmented AI landscape, one in which there can only be losers and no winners.

Navigating ethical dilemmas in Generative AI Development: A framework for responsible innovation

This comprehensive article explores the evolving landscape of ethical challenges and solutions in generative AI development, focusing on key areas, including data privacy, algorithmic bias, and implementation frameworks. Through detailed analysis of case studies and research findings, the study examines successful approaches to ethical AI implementation across various sectors, particularly in content generation and code development. The article presents a structured framework for ethical decision-making, emphasizing the importance of proactive assessment, stakeholder engagement, and continuous monitoring. By investigating cultural integration strategies and technical safeguards, this study provides valuable insights into establishing and maintaining ethical AI systems while balancing innovation with responsible development practices in an increasingly complex technological landscape.

Mathematical and Statistical Tools for Ensuring Fairness in Computational Models

The rising use of computational models in decision-making within healthcare, hiring, and finance underlines that fairness is of pressing concern as a means for ensuring equity, building trust, and promoting accountability. This research systematically investigates mathematical and statistical methods to identify, measure, and reduce biases of the systems. It reviews basic concepts of fairness, such as demographic parity, equality of odds, and individual fairness; methods for detecting and mitigating bias. Techniques discussed range from pre-processing adjustments to in-processing constraints to post-processing optimizations toward trading off between fairness and predictive accuracy. Application cases include the case study on ICU admissions using the MIMIC-III dataset that practically illustrates how effectiveness can be ensured with the help of fairness-aware strategies. Where there was a big gap, it now greatly closes it in demographic parity, reducing disparities in true positive rates; therefore, proving reliability for these models through sensitivity and adversarial testing. This paper bridges the divide between theory and practice by presenting a comprehensive framework towards the consideration of fairness within computational systems. The presented work identifies prospects for enhancing equity while ensuring dependability and supplements a high-growth conversation about ethical practices in AI development.

One does not simply create an AI issue. On the pseudo-problematic nature of AI issue

The debate surrounding the topic of Artificial Intelligence (ai), and its different meanings, seems to be ever-growing. This paper aims to deconstruct the seemingly problematic nature of the ai debate, revealing layers of ambiguity and misperceptions that contribute to a pseudo-problematic narrative. Through a review of existing literature, ethical frameworks, and public discourse, this essay identifies key areas where misconceptions, hyperbole, and exaggerated fears have overshadowed the genuine concerns associated with ai development and deployment. To identify these issues I propose three general criteria that are based on Popper’s and Ayer’s work and adjusted to my needs. The subsequent sections categorize ai issues into ontological, methodological, and logical-grammatical problems, aligning with Cackowski’s typology. In addition, I introduce «» signs to distinguish behavioural descriptions from cognitive states, aiming to maintain clarity between external evidence and internal agent states. My conclusion is quite simple: the ai debate should be thoroughly revised, and we, as scholars, should define the concepts that lie at the bottom of ai by creating a universal terminology and agreeing upon it. This will give us the opportunity to conduct our debates reasonably and understandably for both scholars and the popular public.

External leadership knowledge integration and AI readiness in family firms: exploring the mediation effects of professionalization

PurposeAs the knowledge-based theory of the firm suggests, integrating leadership knowledge from non-family sources may be paramount in building competitive advantage, and leadership knowledge is an invaluable resource in building successful firms. In the volatile business environment, AI developments are among the key drivers of several thorough transformations, so the possible relationship between top management business talent and AI readiness has become essential. Our paper examines this link in the family business context, analyzing the possible mediation role of firm professionalization.Design/methodology/approachThe paper adopted a quantitative research approach. Data about Hungarian family firms were collected during March and April 2024. The authors used structural equation modeling, and results with and without controlling for a firm size and sector effect were compared. The final sample size contained 112 family firms.FindingsThe research findings suggest that the identified personal professionalization subdimensions have different mediation roles in the relationship between external knowledge integration in top management teams and AI readiness. Without including the control variables, the personal development and competence absorption subdimensions fully mediate this relationship, and the delegation subdimension does not have a significant mediation role. These two mediation effects become significantly weaker with the inclusion of firm size and sector control variables.Research limitations/implicationsThe results highlight the importance of competence absorption and development in family firms, especially when preparing for AI integration into business processes. In addition, the findings also emphasize that firm size may have a significant role in creating an AI-compatible business environment.Originality/valueThis paper aims to integrate the evolving field of firm professionalization with research on the effects of family involvement in corporate leadership. Professionalization dimensions are defined, and mediation roles of personal professionalization subdimensions are examined concerning the relationship between external knowledge integration in top management teams and AI readiness. The paper also highlights several new research directions that may enhance understanding the relationship between family leadership involvement and corporate outcomes.

An Empirical Study on Challenges for LLM Application Developers

In recent years, large language models (LLMs) have seen rapid advancements, significantly impacting various fields such as computer vision, natural language processing, and software engineering. These LLMs, exemplified by OpenAI’s ChatGPT, have revolutionized the way we approach language understanding and generation tasks. However, in contrast to traditional software development practices, LLM development introduces new challenges for AI developers in design, implementation, and deployment. These challenges span different areas (such as prompts, APIs, and plugins), requiring developers to navigate unique methodologies and considerations specific to LLM application development. Despite the profound influence of LLMs, to the best of our knowledge, these challenges have not been thoroughly investigated in previous empirical studies. To fill this gap, we present the first comprehensive study on understanding the challenges faced by LLM developers. Specifically, we crawl and analyze 29,057 relevant questions from a popular OpenAI developer forum. We first examine their popularity and difficulty. After manually analyzing 2,364 sampled questions, we construct a taxonomy of challenges faced by LLM developers. Based on this taxonomy, we summarize a set of findings and actionable implications for LLM-related stakeholders, including developers and providers (especially the OpenAI organization).

Recent Advancements in Anomaly Detection for Manufacturing Systems

Recent advancements in anomaly detection and root cause analysis have revolutionized manufacturing processes, enabling improved efficiency, quality control, and cost reduction. This review explores the latest developments in these fields, focusing on supervised and unsupervised anomaly detection methods, statistical techniques for root cause analysis, and predictive modeling approaches. Supervised anomaly detection methods, such as classification-based approaches, support vector machines, and decision trees, leverage labeled data to identify anomalies. Unsupervised techniques, including clustering-based methods (e.g., k-means and DBSCAN), density-based approaches (e.g., local outlier factor), and dimensionality reduction (e.g., PCA and autoencoders), detect anomalies without prior knowledge of anomalous instances. Statistical methods for root cause analysis, such as statistical process control, hypothesis testing, correlation analysis, and multivariate techniques (e.g., factor analysis), enable the identification of the underlying causes of manufacturing issues. Predictive modeling techniques, including time-series forecasting, machine learning for predictive maintenance, and optimization methods for process improvement, further enhance manufacturing efficiency. The integration of these techniques has led to significant improvements in quality control, equipment failure prediction, and supply chain optimization. However, challenges remain in handling complex, high-dimensional data and integrating domain knowledge with data-driven approaches. Future research directions include the development of explainable AI, edge computing, and federated learning for anomaly detection in manufacturing. This analysis offers a thorough examination of cutting-edge techniques for detecting anomalies and analyzing root causes, emphasizing their capacity to transform manufacturing operations and stimulate future advancements in these research areas. Keywords—anomaly detection, manufacturing, supervised and unsupervised anomaly detection

Development of AI Driven website for Predicting and Classifying Various Diseases Using Medical Images

Development of AI Driven website for Predicting and Classifying Various Diseases Using Medical Images

Incremental accumulation of linguistic context in artificial and biological neural networks

Large Language Models (LLMs) have shown success in predicting neural signals associated with narrative processing, but their approach to integrating context over large timescales differs fundamentally from that of the human brain. In this study, we show how the brain, unlike LLMs that process large text windows in parallel, integrates short-term and long-term contextual information through an incremental mechanism. Using fMRI data from 219 participants listening to spoken narratives, we first demonstrate that LLMs predict brain activity effectively only when using short contextual windows of up to a few dozen words. Next, we introduce an alternative LLM-based incremental-context model that combines incoming short-term context with an aggregated, dynamically updated summary of prior context. This model significantly enhances the prediction of neural activity in higher-order regions involved in long-timescale processing. Our findings reveal how the brain’s hierarchical temporal processing mechanisms enable the flexible integration of information over time, providing valuable insights for both cognitive neuroscience and AI development.

Beyond Text Generation: Assessing Large Language Models’ Ability to Reason Logically and Follow Strict Rules

The growing interest in advanced large language models (LLMs) like ChatGPT has sparked debate about how best to use them in various human activities. However, a neglected issue in the debate concerning the applications of LLMs is whether they can reason logically and follow rules in novel contexts, which are critical for our understanding and applications of LLMs. To address this knowledge gap, this study investigates five LLMs (ChatGPT-4o, Claude, Gemini, Meta AI, and Mistral) using word ladder puzzles to assess their logical reasoning and rule-adherence capabilities. Our two-phase methodology involves (1) explicit instructions about word ladder puzzles and rules regarding how to solve the puzzles and then evaluate rule understanding, followed by (2) assessing LLMs’ ability to create and solve word ladder puzzles while adhering to rules. Additionally, we test their ability to implicitly recognize and avoid HIPAA privacy rule violations as an example of a real-world scenario. Our findings reveal that LLMs show a persistent lack of logical reasoning and systematically fail to follow puzzle rules. Furthermore, all LLMs except Claude prioritized task completion (text writing) over ethical considerations in the HIPAA test. Our findings expose critical flaws in LLMs’ reasoning and rule-following capabilities, raising concerns about their reliability in critical tasks requiring strict rule-following and logical reasoning. Therefore, we urge caution when integrating LLMs into critical fields and highlight the need for further research into their capabilities and limitations to ensure responsible AI development.

Piezo Ion Channels and Their Association With Haptic Technology Use: A Narrative Review.

The recent identification of Piezo ion channels demonstrating a mechano-sensitive impact on neurons revealed distinct Piezo-1 and 2 types. While Piezo-1 predominates in neurons linked to non-sensory stimulation, such as pressure in blood vessels, Piezo-2 predominates in neurons linked to sensory stimulation, such as touch. Piezo-1 and 2 have a major bidirectional impact on transient receptor potential (TRP) ion channels, and TRPs also impact neurotransmitter release. Particularly existent in dorsal root ganglion (DRG) neurons, which are located in nerve endings, Piezo-2 is a key DRG activator. Subsequent Piezo findings have been vital to recent medical haptic technology developments, in tandem with breakthroughs in the emerging neurology subfield of connectomics plus AI developments. Included in this review are a historical Piezo overview, the interrelationship of Piezo channels with TRPs, inclusive of TRPV1/TRPV8, the impact on medical and rehab haptic technology, a focus on haptic technology use in stroke survivor rehab inclusive of pain mitigation, and the development of a haptic technology patch aimed at alleviating pain and/or anxiety. Neurogenic pain resulting from hyperalgesia/allodynia in stroke survivors is a potential target for drugs and haptics aimed at pain reduction; patients experiencing neuropathic or psychosomatic pain are other prime targets. Increased Piezo knowledge may promote more precisely targeted haptic therapeutic developments.

Human Formation in the Age of Automation

ABSTRACTThis chapter responds to the recent crisis surrounding developments in large language models (LLMs) and generative AI with a relational view of education informed by the emerging world‐centered approach to education and a synthesis of personalist character formation with feminist care ethics. It proposes that the instinct to manage student use of generative AI with carceral practices, characterized by surveillance and control, undermines the project of forming students as whole persons who can evaluate and use or reject emerging technologies in a mature way. The chapter concludes with practical suggestions for the classroom that can be aligned with institutional strategic plans.

Understanding Agentic Frameworks in AI Development: A Technical Analysis

This technical article examines the evolution and implementation of agentic frameworks in artificial intelligence development, focusing on their transformative impact across multiple industries. The article explores the fundamental architectural components, implementation methodologies, and practical applications of these frameworks in manufacturing, financial services, and healthcare sectors. By investigating the core components, including perception systems and decision architectures, alongside the Belief-Desire-Intention model and advanced learning mechanisms, this article provides comprehensive insights into how agentic frameworks are revolutionizing autonomous decision-making capabilities. The article also addresses critical technical challenges in scalability and safety while offering potential solutions and future directions for development, highlighting the growing importance of these frameworks in shaping the future of AI technology.

Synthetic Data Generation: Enabling Secure Use of Data for AI, Machine Learning, and Testing

This paper explores the critical role of synthetic data generation in enabling secure and privacy-preserving use of data for artificial intelligence, machine learning, and software testing applications. As organizations face increasing regulatory pressures and data privacy concerns, synthetic data emerges as a powerful solution to maintain data utility while mitigating risks associated with sensitive information. We examine the challenges in using real-world data, the need for synthetic data generation, and technical approaches to implementing robust synthetic data solutions. The paper also addresses how synthetic data can enhance AI and ML model development, improve testing processes, and support overall data governance strategies in enterprise environments. This paper explores the challenges organizations face in accessing secure and compliant data, highlights the need for synthetic data, and presents a technical framework for implementing synthetic data generation solutions. Furthermore, it outlines metrics for measuring effectiveness and provides insights into the future potential of synthetic data. Index terms: synthetic data, data privacy, machine learning, artificial intelligence, data security, data augmentation, generative models (key words)

How Can the Protection of Important Agricultural Heritage Sites Contribute to the Green Development of Agriculture: Evidence from China

The protection of agricultural heritage sites has become a global human responsibility and consensus. However, the potential effect of agricultural heritage sites on the green development of agriculture has currently been ignored. Since ancient times, China has been founded on agriculture, and the number of important agricultural cultural heritages ranks first in the world, with strong representativeness. The two-way fixed effects model was employed to empirically test the positive impact of agricultural heritage site protection on the green development of agriculture, utilising data from 30 provincial units in China over a 21-year period from 2001 to 2021 in this paper. Additionally, a mediating effect model was used to test the potential mechanism. The specific conclusions are as follows: firstly, the protection of agricultural heritage sites has a significant positive effect on the green development of agriculture; secondly, the protection of agricultural heritage sites can advance the agricultural industrial structure, increase the land transfer rate, strengthen the construction of new agricultural management organisations, enhance financial investment in supporting, and promote green technology innovation. The five paths were found to have passed the mediation effect test. Thirdly, the positive effect of protecting agricultural heritage sites on the green development of agriculture is heterogeneous. Specifically, the effect is more pronounced in the eastern and central regions, particularly in areas with higher levels of AI development and lower urbanisation. The ultimate objective is to utilise evidence from China to develop nature-based solutions for the protection and utilisation of agricultural heritage and green development of agriculture in other world agricultural heritage sites.

Distribution of responsibility for AI development: expert views

Abstract The purpose of this paper is to increase the understanding of how different types of experts with influence over the development of AI, in this role, reflect upon distribution of forward-looking responsibility for AI development with regard to safety and democracy. Forward-looking responsibility refers to the obligation to see to it that a particular state of affairs materialise. In the context of AI, actors somehow involved in AI development have the potential to guide AI development in a safe and democratic direction. This study is based on qualitative interviews with such actors in different roles at research institutions, private companies, think tanks, consultancy agencies, parliaments, and non-governmental organisations. While the reflections about distribution of responsibility differ among the respondents, one observation is that influence is seen as an important basis for distribution of responsibility. Another observation is that several respondents think of responsibility in terms of what it would entail in concrete measures. By showing how actors involved in AI development reflect on distribution of responsibility, this study contributes to a dialogue between the field of AI governance and the field of AI ethics.

An Emerging Design Space of How Tools Support Collaborations in AI Design and Development

Developing AI/ML models and incorporating them into a product is complex work. While AI models are generally non-deterministic and have high capability uncertainty, the advent of foundation models further exacerbates the complexity of working with AI and ensuring responsible innovation. This complex work is achieved by not just AI practitioners, but through coordination and collaboration of different groups of practitioners, not all of whom might be experts in AI. Our primary objective is to explore how the tools and systems used by practitioners help achieve this complex work. To that end, we conducted a design space analysis of 18 relevant tools using corresponding research publications and constructed a design space with the dimensions - User, Axis of AI work, Semantics of Use, Tool Architecture, Artifact Type and Availability, and Collaboration Goals. Using these dimensions we derive four spirits of the tool in supporting collaborations - groupware, core practice & communication, community of practice, and visibility & bridging. Through this work, we contribute a conceptual design space of how tools can be designed to support collaborations in AI development and discuss how our design space can be leveraged by system designers and researchers working at the intersection of HCI, CSCW, and AI.