In an age of exponential development of neural network technologies for natural language processing, professional and social practices are undergoing unprecedented transformations. Artificial Intelligence (AI) tools, in particular Neural Machine Translation (NMT), can be invaluable resources to assist people in various domains of application, from information access and retrieval to communication exchange and knowledge dissemination (Monti 2019; Koehn 2020). However, alongside the benefits of such instruments, there are also potential risks that require consideration to promote correct, ethical and conscientious use. At a discursive level, NMT technologies can perpetuate stereotypes and prejudices. For example, language processing algorithms can reflect the biases present in their training data, leading to results that reinforce discriminatory social, cultural, and racial representations (see Stanovsky et al. 2019; Marzi 2021). At a professional level, while AI tools can increase efficiency and accuracy, they also require new skills and practices. In professional translation, for example, the development of appropriate competencies is essential to enable translators not only to integrate input from different digital sources, from translation memories to NMT but also to pre- and post-edit texts according to established guidelines (e.g., Translation Automation User Society (TAUS) guidelines, see O’Brien 2022). Finally, at a cultural level, the impact of AI on linguistic diversity and identity needs to be considered. NMT is an example of a neural technology that is increasingly available on major e-commerce platforms, institutional websites, and as a free online application, helping users to overcome language barriers (see Leppänen & Peuronen 2021). While the number of supported languages is increasing, several critical issues need to be addressed, such as the impact of English as a pivot language for deep learning and the scarce availability of large corpora for low-resource languages (Vetere 2022; also see Charlton 2018). 59 In this context, as AI and NMT continue to evolve and become more widespread, it is essential to study and monitor their impact on professional activities, develop strategies to reduce discursive biases, and support and preserve linguistic diversity. Moreover, it is of paramount importance to promote conscious and ethical approaches to the use of AI tools, based on a sound understanding of their implications. By inviting papers that develop critical reflections or case studies on these issues, this panel aims to encourage discussion and reflection on the impact of AI and NMT on social, professional and cultural identities. References Charlton, Emma. 2018. “The Internet has a language diversity problem”. World Economic Forum, 13 December 2018. https://www.weforum.org/agenda/2018/12/chart-of-the-day-the-internet-has-a-language-diversity-problem 2, p. 240-267. Koehn, Philipp. 2020. Neural Machine Translation. Cambridge: Cambridge University Press. Monti, Johanna. 2019. Dalla Zairja alla traduzione automatica. Riflessioni sulla traduzione nell’era digitale. Napoli: Liguori. Leppänen, Sirpa, and Saija Peuronen. 2015. “Multilingualism on the Internet”. In M. Martin-Jones, A. Blackledge, and A. Creese (Eds.), The Routledge Handbook of Multilingualism (pp. 384-403). London/New York: Routledge. Marzi, Eleonora. 2021. “La traduction automatique neuronale et les bais de genre”. Synergies Italie, 17, 19-36. O’Brien, Sharon. 2022. “How to deal with errors in machine translation: Post-editing”. In Kenny, D. (ed.), Machine translation for everyone: Empowering users in the age of artificial intelligence (pp. 105-120). Berlin: Language Science Press. Stanovsky, Gabriel, Noah A. Smith, and Luke Zettlemoyer. 2019. “Evaluating gender bias in machine translation”. In A. Korhonen, D. Traum, L. Màrquez (Eds), Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (pp. 1679–1684). Florence, Italy: Association for Computational Linguistics. Vetere Guido. 2022. “Elaborazione automatica dei linguaggi diversi dall’inglese: introduzione, stato dell’arte e prospettive.” De Europa: European and Global Studies Journal, Special Issue “Multilingualism and Language Varieties in Europe in the Age of Artificial Intelligence”, edited by Rachele Raus, Alida Maria Silletti, Silvia Domenica Zollo and John Humbley, 69–87.

PANEL IV: THE IMPACT OF NMT ON SOCIAL, PROFESSIONAL AND CULTURAL IDENTITIES

Ilaria Cennamo
;
Lucia Cinato
;
Maria Margherita Mattioda
;
Alessandra Molino
2023-01-01

Abstract

In an age of exponential development of neural network technologies for natural language processing, professional and social practices are undergoing unprecedented transformations. Artificial Intelligence (AI) tools, in particular Neural Machine Translation (NMT), can be invaluable resources to assist people in various domains of application, from information access and retrieval to communication exchange and knowledge dissemination (Monti 2019; Koehn 2020). However, alongside the benefits of such instruments, there are also potential risks that require consideration to promote correct, ethical and conscientious use. At a discursive level, NMT technologies can perpetuate stereotypes and prejudices. For example, language processing algorithms can reflect the biases present in their training data, leading to results that reinforce discriminatory social, cultural, and racial representations (see Stanovsky et al. 2019; Marzi 2021). At a professional level, while AI tools can increase efficiency and accuracy, they also require new skills and practices. In professional translation, for example, the development of appropriate competencies is essential to enable translators not only to integrate input from different digital sources, from translation memories to NMT but also to pre- and post-edit texts according to established guidelines (e.g., Translation Automation User Society (TAUS) guidelines, see O’Brien 2022). Finally, at a cultural level, the impact of AI on linguistic diversity and identity needs to be considered. NMT is an example of a neural technology that is increasingly available on major e-commerce platforms, institutional websites, and as a free online application, helping users to overcome language barriers (see Leppänen & Peuronen 2021). While the number of supported languages is increasing, several critical issues need to be addressed, such as the impact of English as a pivot language for deep learning and the scarce availability of large corpora for low-resource languages (Vetere 2022; also see Charlton 2018). 59 In this context, as AI and NMT continue to evolve and become more widespread, it is essential to study and monitor their impact on professional activities, develop strategies to reduce discursive biases, and support and preserve linguistic diversity. Moreover, it is of paramount importance to promote conscious and ethical approaches to the use of AI tools, based on a sound understanding of their implications. By inviting papers that develop critical reflections or case studies on these issues, this panel aims to encourage discussion and reflection on the impact of AI and NMT on social, professional and cultural identities. References Charlton, Emma. 2018. “The Internet has a language diversity problem”. World Economic Forum, 13 December 2018. https://www.weforum.org/agenda/2018/12/chart-of-the-day-the-internet-has-a-language-diversity-problem 2, p. 240-267. Koehn, Philipp. 2020. Neural Machine Translation. Cambridge: Cambridge University Press. Monti, Johanna. 2019. Dalla Zairja alla traduzione automatica. Riflessioni sulla traduzione nell’era digitale. Napoli: Liguori. Leppänen, Sirpa, and Saija Peuronen. 2015. “Multilingualism on the Internet”. In M. Martin-Jones, A. Blackledge, and A. Creese (Eds.), The Routledge Handbook of Multilingualism (pp. 384-403). London/New York: Routledge. Marzi, Eleonora. 2021. “La traduction automatique neuronale et les bais de genre”. Synergies Italie, 17, 19-36. O’Brien, Sharon. 2022. “How to deal with errors in machine translation: Post-editing”. In Kenny, D. (ed.), Machine translation for everyone: Empowering users in the age of artificial intelligence (pp. 105-120). Berlin: Language Science Press. Stanovsky, Gabriel, Noah A. Smith, and Luke Zettlemoyer. 2019. “Evaluating gender bias in machine translation”. In A. Korhonen, D. Traum, L. Màrquez (Eds), Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (pp. 1679–1684). Florence, Italy: Association for Computational Linguistics. Vetere Guido. 2022. “Elaborazione automatica dei linguaggi diversi dall’inglese: introduzione, stato dell’arte e prospettive.” De Europa: European and Global Studies Journal, Special Issue “Multilingualism and Language Varieties in Europe in the Age of Artificial Intelligence”, edited by Rachele Raus, Alida Maria Silletti, Silvia Domenica Zollo and John Humbley, 69–87.
2023
LANGUAGING IDENTITIES IN CHANGING TIMES. Challenges and opportunities
Torino
14-16 dicembre 2023
Languaging Identities in Changing TimesChallenges and opportunities
Collane@unito.it
58
59
9788875902803
https://www.collane.unito.it/oa/items/show/154#?c=0&m=0&s=0&cv=0
Ilaria Cennamo; Lucia Cinato; Maria Margherita Mattioda; Alessandra Molino
File in questo prodotto:
File Dimensione Formato  
4022ed691e2688a302afda8bc9d98de3.pdf

Accesso aperto

Descrizione: Book of abstracts
Tipo di file: PDF EDITORIALE
Dimensione 2.54 MB
Formato Adobe PDF
2.54 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/2318/1946642
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact