ACM launches journal on AI security and privacy
On June 24, 2025, the Association for Computing Machinery (ACM) announced the launch of a new journal, ACM Transactions on AI Security and Privacy (TAISAP), designed to address critical research needs in securing AI systems and leveraging AI for cybersecurity.
The establishment of TAISAP responds to the increasing ubiquity of AI technologies, which has generated a demand for specific research focusing on their security vulnerabilities and defensive measures. Concurrently, the journal will explore how AI can enhance existing cybersecurity frameworks. In line with ACM’s institutional objective to transition all its publications to open access by January 2026, all papers accepted for publication in TAISAP will be made available via open access without any publication charges for an initial period of three years, covering 2026 through 2029.
TAISAP’s primary scope is centered on the development of methodologies for evaluating the security posture of AI models, AI-enabled systems, and the broader environments in which AI operates. This encompasses several key areas: the intrinsic security of AI models, which includes resilience against adversarial attacks and the identification of technical vulnerabilities inherent in AI algorithms; privacy considerations during the training and deployment phases of AI models; and the application of AI and machine learning techniques within cybersecurity domains, such as security operations centers and cyber threat intelligence gathering.
The journal intends to investigate a wide spectrum of research, including technical, behavioral, and economic approaches, to thoroughly assess the security and privacy aspects of AI. This includes detailed examinations of AI models, systems, and their operational environments. TAISAP also aims to publish high-quality scholarly articles that contribute to the development of AI-enabled analytical methods. These methods may leverage machine learning, deep learning, large language models, network science, and other related computational techniques for diverse cybersecurity applications.
TAISAP accepts both theoretical and applied contributions. Theoretical submissions are expected to introduce novel insights concerning threats and corresponding defense mechanisms that impact AI security and privacy within the specified domains. Applied contributions should present innovative methodologies for assessing the security and privacy of AI models and systems. Furthermore, the application of AI in security contexts represents a significant area of focus for the journal. This includes AI-enabled analytics for various cybersecurity functions, such as cyber threat intelligence, vulnerability management, security operations centers, authentication protocols, and open-source software security. The journal encourages multi-disciplinary and interdisciplinary submissions from scholars representing a range of academic fields.
Inside the LLM system that reads emails like a cybersecurity analyst
However, TAISAP’s scope specifically excludes broader socio-technical themes such as AI safety, dependability, and governance. Each submission to the journal is expected to advance the understanding of AI security and privacy in one or more of the aforementioned areas. This can be achieved through the application or development of either practical or theoretical approaches.
The editorial leadership for TAISAP consists of three Co-Editors-in-Chief: Murat Kantarcioglu from Virginia Tech, Patrick McDaniel from the University of Wisconsin, Madison, and Sagar Samtani from Indiana University, Bloomington. The journal’s editorial board also includes 17 Associate Editors, who represent a diverse international cohort from countries including China, Germany, Italy, Liechtenstein, Switzerland, and the United States.
The introduction of TAISAP is part of a larger strategic initiative by ACM to expand its portfolio of journals with a new suite of publications dedicated to various facets of artificial intelligence. TAISAP is expected to commence accepting submissions in the near future. ACM currently publishes more than 70 scholarly peer-reviewed journals, covering numerous disciplines within computing and information technology. These journals form a comprehensive archive of computing innovation, encompassing both emerging and established research for practical and theoretical applications. ACM emphasizes rapid publication to minimize delays in disseminating new ideas and discoveries. ACM, the Association for Computing Machinery, functions as the world’s largest educational and scientific computing society, fostering dialogue, resource sharing, and addressing challenges within the computing field.