Sarah Herbert
About
My research project
AI-enabled accessibility of digital healthcare content for low-resource languagesI am a Leverhulme Trust Doctoral Scholar, researching the role of AI in making digital NHS content accessible across languages. My PhD project follows on from previous research I have done on the use of translation services and technologies across all NHS trusts, councils and police forces England. This previous work uncovered numerous causes for concern with regard to preparation and management of the use of AI, in the form of machine translation (MT) such as Google Translate. Despite finding evidence of MT use in these organisations, there is a lack of regulation and awareness, which can lead to improper and irresponsible practices, with serious consequences for service users, whilst also wasting valuable public resources.
Following from this, my current work looks deeper into the use of MT in digital NHS resources, such as their apps and websites, where my previous project identified extensive use. As part of the AI-enabled Digital Accessibility programme (ADA), this project assesses how these AI tools are being adopted on these platforms and whether they are truly offering the desired accessibility to its users. This study will have a particular focus on speakers of low-resource languages, for which MT is typically less developed. It incorporates the participation of speakers of such languages as well as extensive involvement from NHS stakeholders to assess the effectiveness of these resources and the risks they may pose to patients, staff and the NHS. It will also aim to provide recommendations for improved accessibility, while minimising risks and reducing healthcare inequalities.
Supervisors
I am a Leverhulme Trust Doctoral Scholar, researching the role of AI in making digital NHS content accessible across languages. My PhD project follows on from previous research I have done on the use of translation services and technologies across all NHS trusts, councils and police forces England. This previous work uncovered numerous causes for concern with regard to preparation and management of the use of AI, in the form of machine translation (MT) such as Google Translate. Despite finding evidence of MT use in these organisations, there is a lack of regulation and awareness, which can lead to improper and irresponsible practices, with serious consequences for service users, whilst also wasting valuable public resources.
Following from this, my current work looks deeper into the use of MT in digital NHS resources, such as their apps and websites, where my previous project identified extensive use. As part of the AI-enabled Digital Accessibility programme (ADA), this project assesses how these AI tools are being adopted on these platforms and whether they are truly offering the desired accessibility to its users. This study will have a particular focus on speakers of low-resource languages, for which MT is typically less developed. It incorporates the participation of speakers of such languages as well as extensive involvement from NHS stakeholders to assess the effectiveness of these resources and the risks they may pose to patients, staff and the NHS. It will also aim to provide recommendations for improved accessibility, while minimising risks and reducing healthcare inequalities.
My qualifications
Publications
This article studies the effects of automating a job allocation system, in a translation company of approximately 130 employees. Perceptions of the effects of automation on roles and responsibilities were collected through a short survey, answered by 38 project managers and translators. This evolved to an analysis of effects on the deeper notion of professional responsibility, related to accountability, control, engagement and understanding of a translation workflow. The results first reflect on positive and negative effects of automation, notably indicating that automation can both restrict and enhance professional roles and autonomy. The focus then turns to perceptions of workers’ main responsibilities, when impacted by a new automated process. One key result suggests increased difficulty in prioritising these duties. Furthermore, translators prefer not being restricted by their specialisations and favour the development of new skills. Another relevant finding of the study shows in-house translators as being the group who alludes more frequently to concepts related to responsibility. The article contributes to the study of socio-technical changes in the translation industry, suggesting that responsibility plays an important part in highlighting the effects of technology, not only on professional and organisational practices, but also on individual perceptions of accountability and job satisfaction.
This dataset is composed of a zipped file, which contains five files describing the initial data of the project Translation services & technology in public sector organisations (FOI study). See the readme.txt file for details.
This dataset is composed of a zipped file, which contains seven files describing the data of the project Translation services & technology in public sector organisations (FOI study) relative to the NHS trusts. See the readme.txt file for details.
Over August and November 2023, the Centre for Translation Studies (CTS) sent a Freedom of Information (FOI) request to all NHS trusts in England (210 organisations), as part of the Translation services & technology in public sector organisations (FOI study) - TS&TECC@PSO(FOI) project.The FOI request consisted of questions relating to the use of professional translation and interpreting services, and machine translation (MT) in the NHS, as well as measures taken to ensure the safe and responsible use of such tools. This report outlines some of the key findings of this study.
This 4-part presentation was created to present the Translation services & technology in public sector organisations (FOI study) - TS&TECH@PSO(FOI) project to NHS staff. The presentation begins with case studies that show the risks of AI in multilingual communication, followed by the presentation of the results of the analysis of responses of NHS trusts to the FOI project. The final part discussed potential implications for the NHS of the knowledge generated by the FOI project.
Over August and November 2023, the Centre for Translation Studies (CTS) sent a Freedom of Information (FOI) request to all NHS trusts in England (210 organisations), as part of the Translation services & technology in public sector organisations (FOI study) - TS&TECH@PSO(FOI) project.The FOI request consisted of questions relating to the use of professional translation and interpreting services, and machine translation (MT) in the NHS, as well as measures taken to ensure the safe and responsible use of such tools. This report presents the full analysis of the data produced related to the NHS trusts in this study.
This dataset includes the data collected as part of the projects JAS, a study on the effects of automation of a job allocation system in a translation services company. The dataset includes counts of answers to a questionnaire answered by 38 participants. Answers are classified according to closed classes in closed questions, and thematic codes in answers to open questions. See readme file for more details and read the article "From responsibilities to responsibility: a study of the effects of translation workflow automation", due to be published in JoSTrans, Issue 40 (July 2023).