AI and Ethics, Год журнала: 2024, Номер unknown
Опубликована: Сен. 10, 2024
Язык: Английский
Frontiers in Artificial Intelligence, Год журнала: 2024, Номер 7
Опубликована: Июнь 18, 2024
The release of GPT-4 has garnered widespread attention across various fields, signaling the impending adoption and application Large Language Models (LLMs). However, previous research predominantly focused on technical principles ChatGPT its social impact, overlooking effects human–computer interaction user psychology. This paper explores multifaceted impacts interaction, psychology, society through a literature review. author investigates ChatGPT’s foundation, including Transformer architecture RLHF (Reinforcement Learning from Human Feedback) process, enabling it to generate human-like responses. In terms studies significant improvements GPT models bring conversational interfaces. analysis extends psychological impacts, weighing potential mimic human empathy support learning against risks reduced interpersonal connections. commercial domains, discusses applications in customer service services, highlighting efficiency challenges such as privacy issues. Finally, offers predictions recommendations for future development directions impact relationships.
Язык: Английский
Процитировано
20
Frontiers in Artificial Intelligence, Год журнала: 2025, Номер 7
Опубликована: Янв. 13, 2025
In this article, we introduce a sociolinguistic perspective on language modeling. We claim that models in general are inherently modeling varieties of , and consider how insight can inform the development deployment models. begin by presenting technical definition concept variety as developed sociolinguistics. then discuss could help us better understand five basic challenges modeling: social bias, domain adaptation, alignment, change scale . argue to maximize performance societal value it is important carefully compile training corpora accurately represent specific being modeled, drawing theories, methods, descriptions from field
Язык: Английский
Процитировано
3
British Journal of Educational Technology, Год журнала: 2025, Номер unknown
Опубликована: Апрель 29, 2025
Abstract This paper explores the application of a benefits versus costs reflection approach within non‐university teaching environments, grounded in principles Value‐Sensitive Design. Aimed at integrating human values systematically into adoption digital educational tools, this study involved 136 in‐service school teachers across various workshops Catalonia. Through use structured customisable worksheet, educators critically self‐evaluated their feelings about both and associated with technologies practices. The found that was meaningful to teachers, who were able adapt workshop cases. positive reception by suggests not only satisfactory level usability utility but also agreement need integrate related strategies training, learning design community debate processes. Practitioner notes What is already known topic integration tools education has been widely recognised for its potential enhance experiences, it raises significant ethical practical concerns. Design (VSD) established as framework integrates technology process, particularly human‐computer interaction fields. adds shows VSD form settings, providing new insights perceived usefulness preferences adaptation. It illustrates how can evaluate leading processes enhanced understanding informed debate. Implications practice and/or policy findings advocate inclusion VSD‐based reflective practices teacher training programs educators' ability assess education. Teachers value initiatives implement reflective, participation‐oriented decision‐making aiming maximise minimise harms when selecting technological influence settings.
Язык: Английский
Процитировано
1
Computers and Education Artificial Intelligence, Год журнала: 2024, Номер 7, С. 100272 - 100272
Опубликована: Июль 27, 2024
The increasing significance of artificial intelligence (AI) in various industries highlights the necessity for industry leaders and professionals to comprehend gain knowledge about AI. urgency AI literacy is more critical than ever due potential unethical use resulting from insufficient knowledge. This issue particularly crucial educators because they need understand adapt impact within educational institutions, with some needing become literate AI, given that students now have increased access public tools. research presents a multi-phase study address bridging gap by conducting co-creation innovation champions, illustrated case Generative K-12 school. Using action design research, we engaged experienced teachers who are experts pedagogical co-create generative AI-enhanced platform at leading education institution known its Australia. findings reveal champions enhance their through subject-matter expertise, organizational knowledge, gained external exposure experience. also key elements facilitate cross-domain exchange platform, enabling be exposed experience technological learning, shifts understanding perception initially unaware sceptical aware capable articulating technical rooted shared value. demonstrates how serves as pathway learning among champions. It underscores experiential on collaborative behavioral intentions. Additionally, aligning personal visions values can influence perceptions technologies, enhancing discourse innovation.
Язык: Английский
Процитировано
7
Education and Information Technologies, Год журнала: 2025, Номер unknown
Опубликована: Янв. 3, 2025
Язык: Английский
Процитировано
1
BenchCouncil Transactions on Benchmarks Standards and Evaluations, Год журнала: 2025, Номер unknown, С. 100204 - 100204
Опубликована: Май 1, 2025
Язык: Английский
Процитировано
1
Scientific Reports, Год журнала: 2024, Номер 14(1)
Опубликована: Ноя. 4, 2024
In the financial field, constructing efficient investment portfolios is a focal point of research, encompassing asset selection and optimization allocation. With advancements in Large Language Models (LLMs), generative Artificial Intelligence (AI) tools have showcased capabilities never seen before. However, black-box nature these renders their outputs difficult to interpret directly, often necessitating iterative fine-tuning align with users' expected outcomes. This study presents structured prompt framework specifically designed for stock selection, aiming provide direct interpretable stock-selecting investors various levels. By creating representative scenarios combining them into different cases experimentation, we can explore how construction prompts influences responses generated by AI tools. Additionally, this paper proposes novel algorithm that combines Nonlinear-Activated Beetle Antennae Search strategy Egret Swarm Optimization Algorithm (NBESOA) address Mean-Variance Portfolio Selection problem Transaction Costs Cardinality Constraints (MVPS-TCCC), utilizing real market data construct based on recommendations. Simulation results indicate that, compared other algorithms, NBESOA prefers optimizing portfolio configurations achieve highest Sharpe Ratio strictest constraints, bringing outcomes closer portfolio's frontier.
Язык: Английский
Процитировано
5
Computers and Education Artificial Intelligence, Год журнала: 2024, Номер 7, С. 100329 - 100329
Опубликована: Ноя. 5, 2024
Язык: Английский
Процитировано
5
Learned Publishing, Год журнала: 2024, Номер 38(1)
Опубликована: Дек. 24, 2024
Abstract The aim of this paper is to highlight the situation whereby content generated by large language model ChatGPT appearing in peer‐reviewed papers journals recognized publishers. demonstrates how identify sections that indicate a text fragment was generated, is, entirely created, ChatGPT. To prepare an illustrative compilation appear indexed Web Science and Scopus databases possessing Impact Factor CiteScore indicators, SPAR4SLR method used, which mainly applied systematic literature reviews. Three main findings are presented: highly regarded premier journals, articles bear hallmarks AI models, whose use not declared authors (1); many these identified already receiving citations from other scientific works, also placed found (2); and, most belong disciplines medicine computer science, but there such as environmental engineering, sociology, education, economics management (3). This aims continue add recently initiated discussion on models like creation scholarly works.
Язык: Английский
Процитировано
4
Neural Computing and Applications, Год журнала: 2025, Номер 37(8), С. 6163 - 6179
Опубликована: Янв. 9, 2025
Язык: Английский
Процитировано
0