The Linguist

TheLinguist-64_1-Spring-2025

The Linguist is a languages magazine for professional linguists, translators, interpreters, language professionals, language teachers, trainers, students and academics with articles on translation, interpreting, business, government, technology

Issue link: https://thelinguist.uberflip.com/i/1533066

Contents of this Issue

Navigation

Page 22 of 35

Chartered Institute of Linguists SPRING 2025 The Linguist 23 FEATURES however, clearly misjudged. Pharmacists used it to translate dosing information. Food service assistants used it to ask patients about allergies. Police officers used it to speak to foreign drivers. Balancing risks and benefits In most of these cases, safer communication solutions could have been used. But it is also true that sometimes professional language services are not available. Consider a social worker who has to act on a referral about a child who could be in danger. They need to communicate with Nepali-speaking family members who have low proficiency in English and no knowledge of other languages. The social worker tries to locate a Nepali interpreter but is unable to find one in the time available. A project participant recently described this scenario to me in an interview. They resorted to a mixture of Google Translate and English-speaking family members to communicate. The use of MT in circumstances of this nature is fraught with challenges. For some of the reported problems there was no easy solution. All options had risks as well as benefits. Mistranslations can put lives at risk and so can delays or the wait for a language professional. A concerning finding Uses of MT in the contexts selected for the study involve complex risk-benefit ratios that can change from one minute to the next based on a long list of factors. An outright ban on MT tools is both unfeasible and ill- advised. But what is probably the most concerning finding from the study so far is that MT is not usually mentioned in public- sector workplace training. Its use is common and sometimes necessary but tends to go under the radar. Official guidance on this issue is extremely rare in the UK. Apart from limited statements, or recent considerations on uses of AI more broadly, the questions of language and AI- mediated cross-cultural communication are rarely addressed in policy. This silence is concerning because MT tools are widely available. They are convenient and easy to use, so what may in principle be a last resort translation and interpreting services, provided in a timely fashion by a qualified linguist, will always be safer than a machine translation tool. But among the many factors that need to be considered in this discussion, it is worth noting that sometimes MT is used for purposes that are not currently fulfilled by language professionals. Nurses used it to give directions and help patients find their way around the hospital. Emergency call handlers used it to tell callers that a professional interpreter was being contacted and would be on the line soon. MT here addresses a gap. It does not replace professional linguists. Some uses of MT were, can easily become a first port of call – whether for lack of resources or lack of training and information. Initial findings from this study have recently been published by the Chartered Institute of Linguists in a preliminary report, which makes three basic recommendations: • Organisations to recognise (in training, staff communication and the organisation's literature) that AI/MT exists, and that staff and members of the public may be instinctively inclined to use it. • Institutional policies to address uses of AI in multilingual communication. Policies of this nature need to be flexible enough to keep up with change while also defining standards and the mechanism for assessing needs and for updating the policy itself. • For AI and MT literacy to be embedded in the workplace culture. More emphasis on education and staff training, including when more appropriate solutions, such as professional interpreting, are possible and/or required. The report was not intended to provide detailed best-practice guidance. But hopefully its findings and recommendations will help to raise awareness about an issue that is, for the most part, institutionally ignored in the UK. This issue calls for nuanced discussions concerning not only machine/AI translation and its limitations but also the broader language services ecosystem. Many of the project participants were unhappy with the bureaucracy involved in obtaining professional language assistance. Some complained about telephone interpreting and how sometimes a patchy phone signal, poor sound quality and other seemingly mundane issues got in the way of effective communication. Some complained about the language services they received, which raises questions about standards and the working conditions offered to translators and interpreters. The project is on-going and a full discussion of the findings will be published in due course. In the meantime, a preliminary snapshot of some of the results is available at https://www.ciol.org.uk/ai-translation-uk -public-services. services widely speak its name room

Articles in this issue

Links on this page

Archives of this issue

view archives of The Linguist - TheLinguist-64_1-Spring-2025