This paper highlights the need for robust anonymization techniques to ensure privacy and compliance while maintaining data usability for diverse and critical follow-up tasks, given the increasing prevalence of text data containing sensitive personal information across various fields. This paper provides a comprehensive overview of current trends and recent advances in text anonymization technologies. After discussing foundational approaches, primarily focused on Named Entity Recognition (NAMER), we examine the transformative impact of large-scale language models (LLMs), detailing their dual role as sophisticated anonymization tools and powerful deanonymization threats. We also explore domain-specific challenges and tailored solutions in critical fields such as healthcare, law, finance, and education. We examine advanced methodologies that integrate formal privacy-preserving models with risk-aware frameworks, and address the specialized subfield of author anonymization. Furthermore, we review an evaluation framework, comprehensive metrics, benchmarks, and a practical toolkit for real-world deployment of anonymization solutions. This paper aims to synthesize current knowledge, identify emerging trends and ongoing challenges, including the evolving privacy-utility tradeoff, the need to address quasi-identifiers, and the implications of LLM functionality, and suggest future research directions for both academics and practitioners in this field.
