Trustworthy LLM: AIGC Detection, Alignment and Evaluation

Dear Colleagues,

The rapid advancement of large language models (LLMs) has empowered a new wave of Artificial Intelligence Generated Content (AIGC), enabling transformative applications in education, healthcare, legal services, media, and beyond. With the rapid development of AIGC technologies, ensuring their trustworthiness in terms of security, reliability, and interpretability has become a key issue in the field of artificial intelligence. LLMs and image generation models, although increasingly applied in various domains, present significant risks, including bias, misinformation, the generation of harmful content, and susceptibility to adversarial and backdoor attacks. Recent research has begun to address these challenges by constructing systematic evaluation frameworks for trustworthy LLMs, focusing on dimensions such as model verification and validation, runtime monitoring, alignment evaluation, robustness testing, and security protection.

As a result, trustworthy detection, value alignment, and comprehensive evaluation of large models have become important and timely research frontiers in AI. This Special Issue aims to bring together researchers from academia and industry and present the latest research results in this area. We encourage prospective authors to submit distinguished research papers related to both theoretical approaches and practical case reviews on the subject.

Topics of interest include, but are not limited to, the following:

• Detection and classification of AIGC;
• Alignment techniques for value-consistent model behavior;
• Benchmarking and multi-dimensional evaluation of LLM trustworthiness;
• Defense against hallucination, adversarial inputs, and misuse;
• Auditing, interpretability, and policy-guided generation;
• Cross-modal LLM trust and safety frameworks;
• Application-specific trust analysis in law, healthcare, education, etc.;
• Trustworthiness metrics and benchmarking datasets for LLMs;
• Risk assessment and mitigation of hallucination, bias, and toxicity;
• Formal verification methods for generative models;
• Model auditing, red-teaming, and safety testing tools;
• Privacy-preserving LLM architectures and training pipelines;
• Regulatory compliance and value alignment with legal/ethical norms;
• Human-AI collaboration and feedback-driven alignment;
• Trust calibration and user perception analysis of AIGC systems;
• Watermarking, traceability, and provenance detection for generated content;
• Security threats in multimodal LLMs;
• Model collapse, degradation, and long-term trustworthiness;
• Trustworthy fine-tuning and post-training alignment techniques;
• LLM-based data provenance and model accountability;
• LLM risk governance frameworks for institutional deployment;
• Knowledge-driven Decision Analysis with LLMs;
• Detecting and mitigating jailbreaking and prompt injection.

Through this Special Issue, we aim to promote a deeper understanding of LLM reliability and trust. By highlighting comprehensive solutions, cross-domain case studies, and scalable evaluation frameworks, we aim to bridge existing gaps and foster the development of safe, aligned, and accountable AIGC systems.

Prof. Dr. Li Pan
Dr. Conghui Zheng
Prof. Dr. Ning Liu
Dr. Lifeng Huang
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Electronics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• trustworthy LLM
• AIGC detection
• model alignment
• AI security
• LLM evaluation
• responsible AI