Explore the vast range of titles available.

Generative artificial intelligence (AI) has emerged as a powerful technology with numerous applications in various domains. There is a need to identify the requirements and evaluation metrics for generative AI models designed for specific tasks. The purpose of the research aims to investigate the fundamental aspects of generative AI systems, including their requirements, models, input–output formats, and evaluation metrics. The study addresses key research questions and presents comprehensive insights to guide researchers, developers, and practitioners in the field. Firstly, the requirements necessary for implementing generative AI systems are examined and categorized into three distinct categories: hardware, software, and user experience. Furthermore, the study explores the different types of generative AI models described in the literature by presenting a taxonomy based on architectural characteristics, such as variational autoencoders (VAEs), generative adversarial networks (GANs), diffusion models, transformers, language models, normalizing flow models, and hybrid models. A comprehensive classification of input and output formats used in generative AI systems is also provided. Moreover, the research proposes a classification system based on output types and discusses commonly used evaluation metrics in generative AI. The findings contribute to advancements in the field, enabling researchers, developers, and practitioners to effectively implement and evaluate generative AI models for various applications. The significance of the research lies in understanding that generative AI system requirements are crucial for effective planning, design, and optimal performance. A taxonomy of models aids in selecting suitable options and driving advancements. Classifying input–output formats enables leveraging diverse formats for customized systems, while evaluation metrics establish standardized methods to assess model quality and performance.

Large language models (LLMs) have the potential to transform our lives and work through the content they generate, known as AI-Generated Content (AIGC). To harness this transformation, we need to understand the limitations of LLMs. Here, we investigate the bias of AIGC produced by seven representative LLMs, including ChatGPT and LLaMA. We collect news articles from The New York Times and Reuters, both known for their dedication to provide unbiased news. We then apply each examined LLM to generate news content with headlines of these news articles as prompts, and evaluate the gender and racial biases of the AIGC produced by the LLM by comparing the AIGC and the original news articles. We further analyze the gender bias of each LLM under biased prompts by adding gender-biased messages to prompts constructed from these news headlines. Our study reveals that the AIGC produced by each examined LLM demonstrates substantial gender and racial biases. Moreover, the AIGC generated by each LLM exhibits notable discrimination against females and individuals of the Black race. Among the LLMs, the AIGC generated by ChatGPT demonstrates the lowest level of bias, and ChatGPT is the sole model capable of declining content generation when provided with biased prompts.

The advent of large language models marks a revolutionary breakthrough in artificial intelligence. With the unprecedented scale of training and model parameters, the capability of large language models has been dramatically improved, leading to human-like performances in understanding, language synthesizing, common-sense reasoning, etc. Such a major leap forward in general AI capacity will fundamentally change the pattern of how personalization is conducted. For one thing, it will reform the way of interaction between humans and personalization systems. Instead of being a passive medium of information filtering, like conventional recommender systems and search engines, large language models present the foundation for active user engagement. On top of such a new foundation, users’ requests can be proactively explored, and users’ required information can be delivered in a natural, interactable, and explainable way. For another thing, it will also considerably expand the scope of personalization, making it grow from the sole function of collecting personalized information to the compound function of providing personalized services. By leveraging large language models as a general-purpose interface, the personalization systems may compile user’s requests into plans, calls the functions of external tools (e.g., search engines, calculators, service APIs, etc.) to execute the plans, and integrate the tools’ outputs to complete the end-to-end personalization tasks. Today, large language models are still being rapidly developed, whereas the application in personalization is largely unexplored. Therefore, we consider it to be right the time to review the challenges in personalization and the opportunities to address them with large language models. In particular, we dedicate this perspective paper to the discussion of the following aspects: the development and challenges for the existing personalization system, the newly emerged capabilities of large language models, and the potential ways of making use of large language models for personalization.

Artificial intelligence (AI) holds immense promise for K-12 education, yet understanding the factors influencing students’ engagement with AI courses remains a challenge. This study addresses this gap by extending the technology acceptance model (TAM) to incorporate cognitive factors such as AI intrinsic motivation (AIIM), AI readiness (AIRD), AI confidence (AICF), and AI anxiety (AIAX), alongside human–computer interaction (HCI) elements like user interface (UI), content (C), and learner-interface interactivity (LINT) in the context of using generative AI (GenAI) tools. By including these factors, an expanded model is presented to capture the complexity of student engagement with AI education. To validate the model, 210 Chinese students spanning grades K7 to K9 participated in a 1 month artificial intelligence course. Survey data and structural equation modeling reveal significant relationships between cognitive and HCI factors and perceived usefulness (PU) and ease of use (PEOU). Specifically, AIIM, AIRD, AICF, UI, C, and LINT positively influence PU and PEOU, while AIAX negatively affects both. Furthermore, PU and PEOU significantly predict students’ attitudes toward AI curriculum learning. These findings underscore the importance of considering cognitive and HCI factors in the design and implementation of AI education initiatives. By providing a theoretical foundation and practical insights, this study informs curriculum development and aids educational institutions and businesses in evaluating and optimizing AI4K12 curriculum design and implementation strategies.

This study examines the impact pathway and mediating mechanisms of self-regulated learning (SRL) on AI dependence among university students who frequently use Artificial Intelligence Generated Content (AIGC) tools. A questionnaire survey was conducted among 367 college students in Fujian Province in China. Structural equation modeling analysis revealed that: (1) SRL ability significantly negatively affects AI dependence (β = −.255) and further suppresses AI dependence through the mediating effect of reduced AIGC usage frequency, with the mediation effect accounting for 55.6%; (2) AIGC usage frequency positively exacerbates AI dependence (β = .656), with the highest dependence risk observed in course learning scenarios ( r = .379); (3) Significant group differences exist—female students are more adept at utilizing AIGC for self-directed learning and exhibit lower dependence, while senior students effectively balance tool usage by enhancing their SRL ability. The study reveals the dual inhibitory mechanisms of SRL in technology dependence (direct and mediating effects), emphasizing the need for metacognitive training and differentiated tool usage guidelines tailored to various learning scenarios in educational practice. This provides a theoretical basis for mitigating the risks of AI dependence and ensuring the beneficial application of AIGC tools as an “empowering rather than substitutive” resource.

In an age where computing capabilities are expanding at a breathtaking pace, the advent of Artificial Intelligence-Generated Content (AIGC) technology presents unprecedented opportunities and challenges to the future of design. It is crucial for designers to investigate how to utilize this powerful tool to facilitate innovation effectively. As AIGC technology evolves, it will inevitably shift the expectations of designers, compelling them to delve deeper into the essence of design creativity, transcending traditional sketching or modeling skills. This study provides valuable insights for designers on leveraging AIGC for forward-thinking design innovation. We focus on the representative AIGC tool, “Midjourney”, to explore its integration into design systems for collaborative innovation among content creators. We introduce an AIGC-based Midjourney path for product design and present a supporting tool card set: AMP-Cards. To confirm their utility, we undertook extensive validation through advanced prototype design research, task-specific project practices, and interdisciplinary collaborative seminars. Our findings indicate that AIGC can considerably enhance designers’ efficiency during product development, especially in the “explorative product shape” phase. The technology excels in identifying design styles and quickly producing varied design solutions. Moreover, AIGC’s capacity to swiftly translate creators’ concepts into visual forms greatly aids in multidisciplinary team communication and innovation.

Townscape continuity, which integrates modern architecture with traditional building features, is vital to preserving urban identity. Nevertheless, current architectural practice, particularly façade design, which is the key element in maintaining townscape continuity, remains labor-intensive due to fragmented understanding and inadequate tool support. This study proposes a knowledge-guided AI-generated content (AIGC) approach to improve the efficiency of architectural design. Jiangnan water town in China was adopted as a case study. Specifically, a comprehensive design knowledge repository was developed by integrating academic literature and high-quality design cases. In this process, a large language model (LLM) and a convolutional neural network (CNN) were used to distill the professional achievement, while eye-tracking experiments captured the public perception. This repository was then used to fine-tune a Stable Diffusion model and informed prompt engineering to improve generative quality, thereby improving generative controllability and semantic reliability. A custom evaluation model, trained on preliminary knowledge-guided AIGC, further filtered suboptimal outputs. The final generation results demonstrated strong townscape continuity and adaptability across various architectural scenarios. Finally, expert evaluations confirmed that the proposed approach outperformed baseline models in the design quality. In short, this study offers a systematic knowledge repository for leveraging AIGC in architectural design, supporting practitioners in achieving townscape continuity more effectively.

How human–AI interactive interfaces affect user engagement with AIGC platforms is a critical underexplored issue. Grounded in signaling theory, this paper constructs a theoretical model to examine the differential effects of two interactive interfaces—modular versus open—on user platform engagement, the mediating role of AIGC quality, and the moderating role of user type. Through two scenario experiments, our findings reveal that both AI interactive interfaces (open and modular) significantly enhance user platform engagement, with the open interface exhibiting a stronger effect. Specifically, AIGC accuracy serves as a mediator in the relationship between modular interfaces and user engagement, and innovativeness plays a mediating role in the relationship between the open interface and users’ engagement. Furthermore, novice users strengthen the effect of open AIGC interfaces on AIGC innovation and subsequent engagement, Non-novice users amplify the positive impact of modular AIGC interfaces on both AIGC accuracy; however, there is no significant difference for user engagement. These findings theoretically enrich the customer response model in the literature on human–computer interactions and provide actionable insights for AIGC platform enterprises. By designing tailored interactive interfaces, platforms can generate high-quality, user-perceived AIGC for diverse customer segments. This study offers both theoretical contributions and practical implications for the development of AI-driven user engagement strategies.

[Purpose/Significance] As an emerging technology, the use of artificial intelligence-generated content (AIGC) tools is comprehensively influenced by factors such as individuals, tasks, and tools themselves. From an educational perspective, one effective way to influence user behavior is to improve the outcomes of graduate students' use of AIGC tools. This study aims to reveal the key dimensions and influencing factors of AIGC use by analyzing graduate students' spontaneous behaviors when using AIGC tools. It further seeks to improve the application efficiency of AIGC in graduate students' learning and scientific research, and promote deeper integration between tools and academic activities. [Method/Process] The research follows the logic of \"from the spontaneous behavior of users to the active guidance of educators\", mainly adopting the semi-structured interview method to collect data, and the thematic analysis method to analyze data. Semi-structured interviews were conducted with 25 graduate students from Chinese universities or scientific research institutions. The interviewees included 14 master's students and 11 doctoral students, covering three disciplinary categories: natural sciences (11 students), social sciences (10 students), and humanities (4 students). According to thematic analysis, the interview data were coded, and theoretical saturation was tested. On this basis, a theoretical model of the outcome and its influencing factors of graduate students' use of AIGC tools was constructed, and targeted suggestions were put forward from the perspective of information literacy education. [Results/Conclusions] The use outcome of graduate students' AIGC tool use includes three dimensions: task completion, subjective satisfaction, and process harvest. Its influencing factors involve four aspects: task & situation, personal characteristics, behavioral process, and tool characteristics. 1) task & situation: The use outcome is affected by the matching degree between task demands and application scenarios; 2) personal characteristics: The use outcome is influenced by graduate students' own basic abilities, subjective attitudes, and tool operation skills; 3) behavioral process: The use outcome is significantly impacted by the input of instructions to tools and the provided content; 4) tool characteristics: The use outcome is notably affected by tools' technical functions and operational limitations. Regarding AIGC tool-related education, it is suggested that information literacy educators emphasize the application scenarios of tools, improve the comprehensive ability of graduate students, carry out diversified teaching and training, and pay attention to the dynamics of tool and technology. This study still has some limitations. For instance, it has only identified the dimensions and influencing factors of graduate students' AIGC tool use outcome. Future research will further explore the causal pathways involved in the model through empirical studies.

[Purpose/Significance] In recent years, the rapid development of artificial intelligence (AI) technology has become an international research hotspot. Embedding ChatGPT-like AI system in libraries will lead to a new direction for the development of their intelligent services, which is a necessary choice to improve the library service level, optimize service efficiency and promote service model innovation. It is also an important measure to adapt to the development trend of the information age, which brings new opportunities and challenges to the information and digital construction of libraries. This article aims to explore the realization path of ChatGPT to help libraries provide intelligent services, provide theoretical reference and practical basis for the intelligent service by libraries, and promote the construction of library intelligent service system. [Method/Process] By reviewing the relevant literature at home and abroad, on the basis of summarizing the core technology of ChatGPT, and studying the practical application cases of some domestic libraries, this article analyzes the application scenarios and directions in which ChatGPT can support library services, explains the current limitations of this technology and the threats and challenges it poses to libraries and librarians, and thus proposes corresponding strategies. [Results/Conclusions] ChatGPT technology will further optimize the knowledge service system of libraries and improve the intelligent service capability of libraries. It brings vitality to the libraries, but also brings threats and challenges. Librarians should grasp the integration capabilities of humans and AI in time, and make full use of their unique advantages. Libraries should strengthen the capacity building of their smart services oriented to ChatGPT, innovate the mode and mechanism of their smart services, and rely on ChatGPT to build a smart service platform, provide smart service resources, build a sound service system, build a service team, and improve the level of their smart services. Continuous efforts should be made to promote the in-depth development of smart library services, so as to meet the diversified and personalized information needs of library readers. To this end, this paper makes a preliminary exploration on the opportunities, challenges and development strategies of enabling smart services in libraries through ChatGPT. Due to the limited conditions, only some practices of domestic libraries are explored, and no cases of international libraries are introduced, which has certain limitations and needs to be improved in future studies.