AI-powered translation platform GlobeScribe.ai has launched in the UK, offering fiction translations for US$100 per book per language.

Founders Fred Freeman and Betsy Reavley said the service aims to make fiction translation more accessible and ran blind tests showing readers often couldn’t tell AI from human translations.

Professional translators criticized the quality claims, emphasizing that AI lacks the cultural and emotional sensitivity needed for literary work.

Award-winning translator Deepa Bhasthi highlighted the difficulty AI faces in translating texts from complex languages like Kannada.

She argued that AI tools lack the human understanding necessary to interpret implied and contextual meanings in such texts.

GlobeScribe said its AI tool is not a human replacement but a productivity aid, though concerns persist over its effect on translation quality and the profession.

🔗 Source: The Guardian

🧠 Food for thought

1️⃣ Literary AI translation follows a 75-year evolution with familiar debates

GlobeScribe.ai’s launch represents the latest chapter in machine translation’s long history, which began in 1949 when Warren Weaver first proposed using computers for translation during the Cold War era 1.

The current debate about AI versus human literary translation echoes historical concerns from the 1966 ALPAC report, which criticized early machine translation for failing to match human quality, resulting in reduced funding and research 1.

Translation technology has evolved through distinct phases: rule-based systems in the 1960s-70s, statistical methods in the 1990s, and finally neural machine translation (NMT) in 2013, which revolutionized the field by better capturing context and meaning 2.

Despite these technological advances, the fundamental tension between efficiency and nuance persists. Current AI translation systems achieve only 70-85% accuracy for straightforward content while struggling with idioms and cultural references, compared to the 95-100% accuracy human translators provide 3.

GlobeScribe’s $100 per book pricing reflects this evolution’s economic impact, dramatically undercutting traditional human translation costs that typically range from $500 to $4,000 for a 10,000-word document 3.

2️⃣ Translation industry adapting through hybrid approaches rather than replacement

Despite fears about AI replacing human translators, employment data tells a different story. Translator jobs actually grew by 49.4% between 2008 and 2018, even as translation technology rapidly advanced 4.