Book Review: Davitti,E.,Korybski,T.,& Braun,S. (Eds.),The Routledge Handbook of Interpreting,Technology and AI

The handbook offers a comprehensive single-volume treatment of how technology has reshaped interpreting as a practice, a profession, and a research field. All editors are active researchers in this subfield, leading long-running programmes on video-mediated interpreting and hybrid speech-to-text workflows that recur as evidence across several chapters.

The 22 chapters are distributed across five parts. Part I surveys technology-enabled interpreting, Part II covers technology in interpreter training, Part III treats workflows that automate or semi-automate interpreting, Part IV examines how these technologies land in particular professional settings, and Part V engages cross-cutting debates on quality, ethics, cognition, standards, workflows, and accessibility.

Contributors include practitioner-researchers, computational linguists, and interpreting scholars, with the empirical base predominantly European and more modest engagement with North American legal settings, Chinese-language scholarship, signed-language interpreting, and accessibility for blind interpreters. The trajectory mirrors how the field itself has matured, moving from describing what the tools can do towards examining what they actually do to the practice, the profession, and the people who depend on it.

Chapters under Part I survey the modalities through which technology mediates interpreting, moving from the oldest to the newest. Lázaro Gutiérrez (Chapter 1) covers telephone interpreting and its emergency and public-service histories, with attention to the absence of visual context as both a constraint and a working condition. Braun (Chapter 2) provides the conceptual framework that distinguishes video remote from videoconference interpreting and surveys its empirical study. Chmiel and Spinolo (Chapter 3) examine remote simultaneous interpreting, with experimental work on sound quality, boothmate presence, and cognitive load. Warnicke (Chapter 4) treats video relay service for signed-language interpreting and the interactional dynamics of the configuration in which only the interpreter has access to both channels. Korybski (Chapter 5) makes the historical and practical case for portable interpreting equipment. Ünlü (Chapter 6) reviews automatic speech recognition (ASR)-assisted consecutive interpreting and engages Chinese-language research that English-language surveys often miss. Saina (Chapter 7) closes Part I with tablet interpreting and its situated and extended-cognition framings. Together, the seven chapters offer comprehensive coverage of technology-enabled interpreting and form a valuable reference for students, lecturers, and early researchers. Giving signed-language interpreting its own chapter is one of the volume’s strengths and sets it apart from many comparable surveys. Some overlap is inevitable in such a coordinated treatment, and a few definitional and conceptual threads appear across chapters, particularly around video-mediated configurations, ASR-assisted consecutive workflows, and cognitive load in remote simultaneous interpreting. One area where the part could have gone further concerns the platform on which most current Remote Simultaneous Interpreting (RSI) is delivered. Zoom is identified across several chapters as the dominant platform in practice, but the volume does not engage the growing body of empirical work on ASR-generated live captioning as an interpreting aid. This includes Zoom-specific eye-tracking and interview studies (Yuan & Wang, 2023), the comparative work on ASR captions for accented speech (Cheung & Li, 2022), and the recent triangulated investigations using accuracy measures, NASA Task Load Index (NASA-TLX), eye-tracking, and electroencephalography (EEG; Li & Chmiel, 2024). Given the platform’s centrality to current professional practice, a sustained empirical engagement with it would have strengthened Part I further.

Part II gathers three chapters that together provide a useful map of the field. Prandi (Chapter 8) imposes a clear generational taxonomy on the field of computer-assisted interpreting (CAI) tools and reviews the empirical work with self-criticism, flagging that system-performance benchmarks are often produced by the tools’ own developers. The chapter sits somewhat awkwardly in a part on training, with relatively little on pedagogy itself. Orlando (Chapter 9) makes a conceptual case for smartpens, observing that synchronised playback renders the note-taking process observable for teaching. Amato, Russo, Carioli, and Spinolo (Chapter 10) offer a useful three-wave history of computer-assisted interpreter training and a clarifying distinction between purpose-built training tools and imported professional platforms. The account of immersive environments centres on the avatar-based interpreting in virtual reality (IVY) and evaluating the education of interpreters and their clients through virtual learning activities (EVIVA) projects. It could have been extended to the more recent turn towards fully immersive virtual reality, including the Monash virtual-reality training project for community interpreters (Gerber et al., 2021). A characteristic of the part, as of much of this fast-moving field, is that several contributors assess tools or programmes they have themselves built, and that the tools are evaluated chiefly for usability and user satisfaction. Whether they demonstrably improve interpreting performance remains, by the authors’ own admission, largely open.

Part III, “Technology for (semi-)automating interpreting workflows,” is one of the volume’s most forward-looking sections and addresses what is arguably the fastest-moving area of the field. It comprises two chapters that work well together. Davitti (Chapter 11) brings real-time speech-to-text practices such as live captioning into the scope of interpreting studies, arguing that they share with traditional interpreting the defining feature of immediacy. This shift lets her bring practices such as live captioning and real-time subtitling into the scope of interpreting studies, since these are real-time even though the output is written. The chapter’s central contribution is a continuum of five workflows for live interlingual speech-to-text, ranging from human-centric interlingual respeaking at one end to fully automated ASR-plus-machine-translation at the other. This framework is original and useful. The chapter is also honest about the methodological challenges in the field, particularly the absence of standardised, comparable quality assessment. Several other chapters in the volume raise the same concern. Fantinuoli (Chapter 12) takes the automated end of Davitti’s continuum and explains how it works, distinguishing speech-to-speech translation from machine interpreting and surveying cascading versus end-to-end architectures. The two chapters complement each other well. Part III covers the area of the field that is changing most rapidly, and two further topics would have fitted it naturally. The first is the automation of signed-language interpreting, including speech-to-sign pipelines and signing-avatar synthesis developed in European Union (EU)-funded projects such as SignON (Shterionov et al., 2021, 2024), and engaged with Deaf-community concerns about co-creation and acceptance (O’Boyle et al., 2024). The second is automated quality estimation, which Fantinuoli identifies as an active concern in current machine-interpreting workflows (Kocmi & Federmann, 2023). The volume gives this topic a full chapter in Chapter 17, although as a methodological debate in Part V rather than as a workflow component in Part III. Both framings are defensible, but the workflow framing has become increasingly prominent in the broader speech-translation research community, as evidenced by quality estimation’s inclusion as a dedicated Metrics track at the International Conference on Spoken Language Translation (IWSLT) 2026.

Part IV takes the technologies surveyed earlier and asks how they land in particular professional contexts. These cover conference, health care, legal, and asylum and refugee settings, all high-stakes environments where the consequences of poor communication are severe and full automation is unlikely to be appropriate. Seeber’s chapter on conference settings (Chapter 13) uses a five-component task model to ask which component each technology actually affects. He shows, for instance, that digital voice recorders and digital pens change the consecutive task into something closer to simultaneous interpreting rather than simply making it easier. De Boe’s chapter on health care (Chapter 14) takes an institutional view, situating health care interpreting within the wider digitalisation of health care and resisting the assumption that video-mediated interpreting (VMI) is straightforwardly replacing telephone interpreting (TI). Section 14.4 raises the equity dimension that MT use may reinforce social inequalities by failing to cover languages most relevant for migrant health care. Devaux’s chapter on legal settings (Chapter 15) is most useful on how technology reshapes the courtroom physically and on the ethical implications, where his argument that codes of ethics may not adequately address technology-mediated dilemmas is a substantive contribution. Singureanu and Braun’s chapter on asylum and refugee settings (Chapter 16) draws on the ongoing EU Web-Based Public Service Interpreting (EU-WEBPSI) project to show that technology-mediated interpreting in asylum contexts has measurable consequences for legal outcomes and to propose minimum VMI standards. The four chapters share a common backbone of studies and a shared observation that the spread of these technologies in high-stakes settings is driven less by their technical merits than by institutional and financial conditions, and that the field has yet to develop the quality assessment instruments these settings require.

Part V opens with the chapter on quality-related aspects (Chapter 17) written by Elena Davitti, Tomasz Korybski, Constantin Orăsan, and Sabine Braun. The volume offers a thorough treatment of how quality should be defined, measured, and assessed in technology-mediated interpreting. The chapter’s conceptual framing is properly multidimensional, and automated methods offer a useful methodological map. One observation that might invite future work is that the assessment methods surveyed focus mainly on faithfulness to the source, with less developed treatment of other dimensions such as delivery and target-language quality. Multimodality is touched on but could be more fully woven into the assessment frameworks, given that other chapters in the volume treat it as central to interpreting. The growing role of CAI tools and machine support in the interpreting workflow also raises questions about how traditional assessment rubrics should be weighted, which the chapter acknowledges but does not develop. Giustini’s chapter on ethical aspects (Chapter 18) is the volume’s most direct engagement with the labour, political-economic, and decolonial dimensions of interpreting technology. It covers working conditions, confidentiality, accessibility and bias, and crisis-prone settings such as asylum proceedings, where artificial intelligence (AI)-powered translation has reportedly contributed to asylum rejections. These concerns are increasingly supported by survey evidence on the consequential risks of machine translation (MT) errors in high-stakes settings (Wang, 2025). Mellinger (Chapter 19) maps cognitive-science vocabulary onto interpreting technology and argues for 4EA (embodied, embedded, enactive, extended, and affective) and augmented cognition as frameworks for understanding interpreters who work with tools such as ASR and MT. Pérez Guarnieri and Ghinos (Chapter 20) provide a detailed account of how ISO standards for interpreting are developed and what each existing standard covers. The chapter is especially valuable for its three-layer model of conference interpreting service in ISO 23155, its observation that technical ISO standards contain extensive non-technical provisions about quality and working conditions, and its “four degrees of separation” model linking distance interpreting to cognitive load. Rütten (Chapter 21) is the most practitioner-grounded contribution, organising the analysis around a central thesis of progressive “intensification” of interpreter work, with concrete operational details that will be useful for practitioners. Figiel (Chapter 22) addresses the accessibility of interpreting technology itself, written by a conference interpreter who is visually impaired and who draws on his lived experience and his work with blind students. The chapter surveys specific CAI tools, simultaneous interpreting delivery platforms (SIDPs), and teaching platforms for accessibility and situates this discussion in the human-rights regulatory landscape, including the UN Convention on the Rights of Persons with Disabilities, the EU Web Accessibility Directive, and the European Accessibility Act. Some of the territory these six chapters cover overlaps with material in earlier parts, but the reframing is productive and the closing ties the volume’s concerns back to the conditions under which interpreters actually work and to the people for whom and by whom these technologies are used.

The handbook enters this debate with a timely, interdisciplinary perspective, drawing together cognitive, methodological, ethical, and operational threads at a moment of rapid change. I recommend it to researchers, practitioners, postgraduate students, and trainers at the intersection of interpreting, technology, and AI and to anyone seeking to understand how the profession is being reshaped.