In recent years, the translation profession has encountered unprecedented challenges as artificial intelligence (AI) technologies increasingly handle text and speech translation tasks almost instantaneously. Individuals specializing in less common languages, such as Timothy McKeon, an Irish-language translator, have seen decades of steady employment opportunities with European Union institutions significantly diminish. McKeon reports losing approximately 70% of his income due to a steep drop in available EU translation projects. The advent and proliferation of AI translation software have drastically altered the employment landscape for many linguists worldwide.
Rather than engaging in traditional translation, current work mainly involves revising and refining machine-generated outputs. However, McKeon rejects these assignments on principle, viewing them as complicit in perpetuating technology that increasingly displaces human translators. He explains that by correcting AI translations, translators inadvertently contribute to the learning algorithms of the software, which assimilate these edits to enhance future performance. This process accelerates human translators' obsolescence, equating to 'digging your own professional grave' as human expertise becomes less required.
The translation industry is no longer confronting a hypothetical AI impact—it is experiencing it firsthand. Language apps like Google Translate have been instrumental in reducing the necessity for professional human translators over the past decade. Subsequently, the integration of generative AI capable of producing sophisticated text, images, and audio from simple prompts has intensified this trend. A 2024 survey by the United Kingdom's Society of Authors focusing on writing professionals reveals over one-third of translators have suffered job losses attributable to generative AI. Further, 43% noted a decline in income directly linked to these technological advances.
Research conducted in the United States between 2010 and 2023 by Oxford University's Carl Frey and Pedro Llanos-Paredes examined regional adoption of Google Translate and its effects on translator job growth. Their analysis suggests regions with higher Google Translate users experienced slower increases in professional translation positions. Initially employing statistical translation models, Google Translate transitioned to neural machine translation in 2016, yielding more fluent and natural textual conversions. Frey estimates that approximately 28,000 additional translator jobs might have existed absent machine translation technologies. While this does not yet represent widespread displacement, he anticipates such consequences may emerge.
Globally, translators appear to face similar pressures. McKeon is involved with the Guerrilla Media Collective, comprising translators and communications experts from various countries, all experiencing income supplementation through alternative employment due to AI's impact on their profession.
Within the United States, the situation has sparked legislative debate, notably in Wisconsin. Christina Green, president of Green Linguistics and a court interpreter in that state, expresses concern about proposed legislation permitting courts to utilize AI or machine translation tools in civil, criminal, and other proceedings. This bill, introduced in May, has faced opposition from language professionals who fear the erosion of interpreter demand. Green remarks, "The entire US is looking at Wisconsin" as a potential precedent. While she currently retains court assignments, her company recently lost a substantial Fortune 10 client that switched to an AI-based translation service. This client constituted a significant portion of her business, prompting unavoidable staff reductions.
Green criticizes the general misunderstanding of AI among clients, noting, "People and companies think they're saving money with AI, but they have absolutely no clue what it is, how privacy is affected, and what the ramifications are." She underscores government bodies' inadequate actions in safeguarding privacy and industry standards.
London-based Arabic translator and interpreter Fardous Bahbouh observes a considerable contraction of written translation assignments, attributing it both to technological innovations and media organizations' financial constraints. Pursuing a doctoral degree examining the translation sector, Bahbouh stresses technology's substantial influence on translation and interpretation professions. She calls for increased governmental support to aid affected workers in transitioning to alternative employment avenues to avoid exacerbating inequalities and in-work poverty, pointing out risks such as rising child poverty.
Ian Giles, chair of the Translators Association at the UK's Society of Authors and a professional translator himself, confirms that diminishing incomes are driving many in the field to retrain for other careers. Similar trends are evident in the U.S., where Andrew Benzo, president of the American Translators Association, reports numerous translators abandoning the profession. Supporting this narrative, Kristalina Georgieva, managing director of the International Monetary Fund, disclosed at the World Economic Forum in Davos that the Fund reduced its official translators and interpreters workforce from 200 to 50, largely due to increased reliance on technology.
Bahbouh advocates for enhanced labor protections for those who remain in translation roles to mitigate precarious employment conditions.
Despite rapid advancements in AI-based translation and interpretation, human language professionals retain a crucial role in various high-stakes environments. Benzo highlights the "humongous" risks associated with deploying AI translations in legal, diplomatic, financial, and medical sectors. She emphasizes that precise word choice and subtle nuances inherent in legal or linguistic contexts remain beyond current large language models' capabilities. Translations in these domains demand the expertise and judgment only a skilled human can provide.
While AI integration often suits low-risk daily uses like navigation assistance, critical communications continue to require human oversight.
Literary translation, too, remains a domain relatively shielded from machine substitution. Giles, who translates Scandinavian commercial fiction into English, has lost supplemental commercial translation work owing to AI competition but continues to receive commissions for literary projects.
Finally, Carl Frey points to an essential dimension that AI cannot replicate: interpersonal connection. Pervasive machine translation does not eliminate the need to learn and speak foreign languages for building meaningful relationships across cultures. This human element serves as a persistent requirement irrespective of technological progress.