Introduction to Generative AI in Cybersecurity DOI
Azeem Khan, N. Z. Jhanjhi,

Ghassan A. A. Abdulhabeb

et al.

Advances in human and social aspects of technology book series, Journal Year: 2024, Volume and Issue: unknown, P. 1 - 44

Published: Sept. 13, 2024

The intent of this chapter is to introduce the reader foundations upon which Generative Artificial Intelligence (GenAI) slowly revolutionizing field cybersecurity. Over next few pages, will become familiar with concept GenAI, including its core technologies-generative adversarial networks, variational autoencoders, and a host other sophisticated deep learning models. One needs note that most technologies mentioned in are among cutting-edge developments currently pushing boundaries More importantly, works discussed explain how GenAI allows for new methods identification, detection, prediction, mitigation cyber threats. Nonetheless, tale cybersecurity tackled mixed emotions. Despite enormous promise it holds securing our digital habitats, technology exposed world dangers, especially form privacy invasion potential malpractice. Thus, as defensive tool an offensive weapon, calling balanced strategy govern adoption. Further, should use on understand role interdisciplinary cooperation ethical guidelines address downsides applications. By blending insights revelations from academic practical standpoint, highlighted can change face apart implications emphasizes significance equipping professionals knowledge technologies, advocating proactive adaptable security posture within organizations, well pivotal ongoing research policy development dynamic field. In conclusion, looks into future AI-driven era highlighting sustained innovation, consideration, collaborative efforts ensure landscape evolves by incorporating generative AI advancements.

Language: Английский

The Inadequacy of Reinforcement Learning From Human Feedback—Radicalizing Large Language Models via Semantic Vulnerabilities DOI
Timothy R. McIntosh, Teo Sušnjak, Tong Liu

et al.

IEEE Transactions on Cognitive and Developmental Systems, Journal Year: 2024, Volume and Issue: 16(4), P. 1561 - 1574

Published: March 18, 2024

This study is an empirical investigation into the semantic vulnerabilities of four popular pre-trained commercial Large Language Models (LLMs) to ideological manipulation. Using tactics reminiscent human conditioning in psychology, we have induced and assessed misalignments their retention LLMs, response 30 controversial questions that spanned a broad social spectrum, encompassing both extreme left-wing right-wing viewpoints. Such arise due fundamental limitations LLMs' capability comprehend detailed linguistic variations, making them susceptible manipulation through targeted exploits. We observed xmlns:xlink="http://www.w3.org/1999/xlink">Reinforcement Learning from Human Feedback (RLHF) effect LLM initial answers, but highlighted RLHF two aspects: (1) its inability fully mitigate impact prompts, leading partial alleviation vulnerabilities; (2) inadequacy representing diverse set "human values", often reflecting predefined values certain groups controlling LLMs. Our findings provided evidence inherent current challenged robustness adequacy as mainstream method for aligning LLMs with values, underscored need multidisciplinary approach developing ethical resilient xmlns:xlink="http://www.w3.org/1999/xlink">Artificial Intelligence (AI).

Language: Английский

Citations

31
Creativity and Machine Learning: A Survey DOI Open Access
Giorgio Franceschelli, Mirco Musolesi

ACM Computing Surveys, Journal Year: 2024, Volume and Issue: 56(11), P. 1 - 41

Published: May 11, 2024

There is a growing interest in the area of machine learning and creativity. This survey presents an overview history state art computational creativity theories, key techniques (including generative deep learning), corresponding automatic evaluation methods. After presenting critical discussion contributions this area, we outline current research challenges emerging opportunities field.

Language: Английский

Citations

8
How Can the Current State of AI Guide Future Conversations of General Intelligence? DOI Creative Commons
Tomoe Kanaya,

Ali Magine

Journal of Intelligence, Journal Year: 2024, Volume and Issue: 12(3), P. 36 - 36

Published: March 20, 2024

Similar to the field of human intelligence, artificial intelligence (AI) has experienced a long history advances and controversies regarding its definition, assessment, application. Starting over 70 years ago, AI set out achieve single, general-purpose technology that could overcome many tasks in similar fashion humans. However, until recently, implementations were based on narrowly defined tasks, making systems inapplicable even slight variations same task. With recent towards more generality, contemplation general (AGI) akin (HGI) can no longer be easily dismissed. We follow this line inquiry outline some key questions conceptual challenges must addressed order integrate AGI HGI enable future progress unified intelligence.

Language: Английский

Citations

5
An Investigation into the Utility of Large Language Models in Geotechnical Education and Problem Solving DOI Creative Commons
Liuxin Chen, Amir Tophel, Umidu Hettiyadura

et al.

Geotechnics, Journal Year: 2024, Volume and Issue: 4(2), P. 470 - 498

Published: May 9, 2024

The study explores the capabilities of large language models (LLMs), particularly GPT-4, in understanding and solving geotechnical problems, a specialised area that has not been extensively examined previous research. Employing question bank obtained from commonly used textbook engineering, research assesses GPT-4’s performance across various topics cognitive complexity levels, utilising different prompting strategies like zero-shot learning, chain-of-thought (CoT) prompting, custom instructional prompting. reveals while GPT-4 demonstrates significant potential addressing fundamental concepts its effectiveness varies with specific topics, task, employed. paper categorises errors encountered by into conceptual, grounding, calculation, model inherent deficiencies related to interpretation visual information. Custom prompts, specifically tailored address shortcomings, significantly enhance performance. achieved an overall problem-solving accuracy 67% higher than 28.9% learning 34% CoT. However, underscores importance human oversight interpreting verifying outputs, especially complex, higher-order tasks. findings contribute limitations current LLMs educational fields, providing insights for educators researchers integrating AI tools their teaching approaches. advocates balanced integration education enrich delivery experience emphasising indispensable role expertise alongside technological advancements.

Language: Английский

Citations

4
Open Science at the Generative AI Turn: An Exploratory Analysis of Challenges and Opportunities DOI
Mohammad Hosseini, Serge P. J. M. Horbach, Kristi Holmes

et al.

Published: May 24, 2024

Technology influences Open Science (OS) practices, because conducting science in transparent, accessible, and participatory ways requires tools/platforms for collaborative research sharing results. Due to this direct relationship, characteristics of employed technologies directly impact OS objectives. Generative Artificial Intelligence (GenAI) models are increasingly used by researchers tasks such as text refining, code generation/editing, reviewing literature, data curation/analysis. GenAI promises substantial efficiency gains but is currently fraught with limitations that could negatively core values fairness, transparency integrity, harm various social actors.In paper, we explore possible positive negative impacts on OS. We use the taxonomy within UNESCO Recommendation systematically intersection conclude using advance key objectives further broadening meaningful access knowledge, enabling efficient infrastructure, improving engagement societal actors, enhancing dialogue among knowledge systems. However, due limitations, it also compromise equity, reproducibility, reliability research, while having potential implications political economy its infrastructure. Hence, sufficient checks, validation critical assessments essential when incorporating into workflows.

Language: Английский

Citations

4
Generation Z: AI Affinity and Adoption in Competitive German Organisations DOI Creative Commons

Tobias Nebgen,

Walter Kurz

Published: Jan. 21, 2025

Generation Z, having matured in an entirely digital environment, plays a central role the adoption of AI within organisations. presents potential advantages such as enhanced productivity, process optimisation, and novel employment sectors, while simultaneously posing risks including job displacement, algorithmic biases, ethical dilemmas. This paper examines opportunities challenges associated with this development. The study incorporates literature review two surveys conducted among LinkedIn users across diverse industries to assess Z implementation relevance AI-based systems for competitiveness. Data was collected over seven-day period December 2024. first survey, comprising 202 participants (n = 202), focused on integration use companies. second involving 345 respondents 345), explored whether companies can remain competitive next three five years without AI-supported systems. A target function developed formalise business success context integration, considering key factors technology acceptance, training intensity, workplace design. findings indicate that 58.42% consider contributors total 69.57% indicated they believe German maintain their competitiveness AI, whereas 30.43% regarded critical maintaining While exhibits high level technological affinity, older generations demonstrate more cautious approach adoption. elucidates is contingent upon balance between acceptance supportive measures transparent system results important To address social psychological concerns, insecurity cognitive strain, should adopt structured training, mentoring programmes, change management support responsible integration. formal model implies flexible design organisational culture innovation contribute successful implementation.

Language: Английский

Citations

0
Enabling people-centric climate action using human-in-the-loop artificial intelligence: a review DOI Creative Commons
Ramit Debnath, Nataliya Tkachenko, Malay Bhattacharyya

et al.

Current Opinion in Behavioral Sciences, Journal Year: 2025, Volume and Issue: 61, P. 101482 - 101482

Published: Jan. 24, 2025

Language: Английский

Citations

0
Revolutionizing oncology: the role of Artificial Intelligence (AI) as an antibody design, and optimization tools DOI Creative Commons
Varun Dewaker, Vivek Kumar Morya, Yeon-Ju Kim

et al.

Biomarker Research, Journal Year: 2025, Volume and Issue: 13(1)

Published: March 29, 2025

Antibodies play a crucial role in defending the human body against diseases, including life-threatening conditions like cancer. They mediate immune responses foreign antigens and, some cases, self-antigens. Over time, antibody-based technologies have evolved from monoclonal antibodies (mAbs) to chimeric antigen receptor T cells (CAR-T cells), significantly impacting biotechnology, diagnostics, and therapeutics. Although these advancements enhanced therapeutic interventions, integration of artificial intelligence (AI) is revolutionizing antibody design optimization. This review explores recent AI advancements, large language models (LLMs), diffusion models, generative AI-based applications, which transformed discovery by accelerating de novo generation, enhancing response precision, optimizing efficacy. Through advanced data analysis, enables prediction sequences, 3D structures, complementarity-determining regions (CDRs), paratopes, epitopes, antigen-antibody interactions. These AI-powered innovations address longstanding challenges development, improving speed, specificity, accuracy design. By integrating computational with biomedical driving next-generation cancer therapies, transforming precision medicine, patient outcomes.

Language: Английский

Citations

0
Human, All Too Human: A Philosophical Investigation on Intellectual Property Rights for AI-Based Creativity DOI
Antonino Rotolo

Published: Jan. 1, 2025

Language: Английский

Citations

0
How are Machine Learning and Artificial Intelligence Used in Digital Behavior Change Interventions? A Scoping Review DOI
Amy Bucher, E. Susanne Blazek,

Christopher T. Symons

et al.

Mayo Clinic Proceedings Digital Health, Journal Year: 2024, Volume and Issue: 2(3), P. 375 - 404

Published: May 22, 2024

Language: Английский

Citations

3