Latest news with #AITranslation
The Guardian
08-07-2025
- Business
- The Guardian
AI translation service launched for fiction writers and publishers prompts dismay among translators
An AI fiction translation service aimed at both traditional publishers and self-published authors has been launched in the UK. is currently charging $100 per book, per language for use of its translation services. 'There will always be a place for expert human translation, especially for highly literary or complex texts,' said the founders Fred Freeman and Betsy Reavley, who previously founded Bloodhound Books, which specialises in crime and thrillers. 'But opens the door to new opportunities, making translation a viable option for a much broader range of fiction.' GlobeScribe conducted 'extensive blind testing' of its tool. Native speakers reviewed GlobeScribe translations alongside human-translated versions of texts without being told which method had been used. 'The feedback consistently showed that readers could not reliably distinguish between them,' according to a company statement. 'In some cases, reviewers even felt the AI-assisted versions were closer in tone and fidelity to the original English manuscript.' However, prominent translators along with a translators' organisation have expressed concern over the initiative. GlobeScribe 'may claim to unlock global access for fiction, but their approach sidelines the very people who make literature resonate across cultures,' said Ian Giles, chair of the Society of Authors' Translators Association. 'Suggesting that AI can match, or even surpass, the nuanced work of human translators on behalf of authors is flat-out wrong.' 'The best literary translations offer more than simple accuracy, more than literal fidelity to the words making up the sentences,' said Polly Barton, writer and translator of works including the bestselling Butter by Asako Yuzuki from Japanese to English. 'They are engaging with the context from which the book has come, and reproducing the pacing, atmosphere, emotional timbre, rhythm, and all the other, less superficially obvious factors that ultimately determine how fulfilling and rich the reading experience is.' Deepa Bhasthi, whose translation of Banu Mushtaq's Heart Lamp from Kannada into English won this year's International Booker prize, said that 'there are many words in Kannada that contain whole cultural worlds in them, where there is as much hidden or implied within a cultural context as is conveyed in literal meaning. And that needs a human being, with an understanding of these visible and invisible worlds, to translate such words'. Asked about GlobeScribe's testing method, Barton said that being a native speaker of a language 'doesn't necessarily equip one to judge all translations from that language with total authority'. Bhasthi added that 'we are not told what kind of texts they were given, what kind of readers the test subjects were'. GlobeScribe's founders said that while they 'recognise that parts of the industry are understandably cautious about what AI might mean for the arts', they 'believe these tools are here to stay and that they should be embraced thoughtfully and responsibly'. They added that AI could enhance creativity and help professional translators 'increase their productivity and output'. The founders 'are clear that this is not about replacing human translators'. Julia Sanches – the translator of works including Boulder by Eva Baltasar from Catalan into English – said: 'Even though I don't think Globescribe can translate the kinds of literary texts I translate, I am gloomy about the emergence of all these new AI 'translation' services. They give the appearance that translation is instant, which devalues my labour, and also that it is mediocre, which could make 'good enough' the new standard for the literary arts. And that's a disservice to both authors and readers.' 'This doesn't just end with translation,' said Barton. 'Maybe translators are at the frontline of people being put out of work by AI technologies, but soon there will be more and more jobs threatened to be wiped out in this way. It lies in our hands to decide whether or not we want this to happen.'
The Guardian
08-07-2025
- Business
- The Guardian
AI translation service launched for fiction writers and publishers prompts dismay among translators
An AI fiction translation service aimed at both traditional publishers and self-published authors has been launched in the UK. is currently charging $100 per book, per language for use of its translation services. 'There will always be a place for expert human translation, especially for highly literary or complex texts,' said the founders Fred Freeman and Betsy Reavley, who previously founded Bloodhound Books, which specialises in crime and thrillers. 'But opens the door to new opportunities, making translation a viable option for a much broader range of fiction.' GlobeScribe conducted 'extensive blind testing' of its tool. Native speakers reviewed GlobeScribe translations alongside human-translated versions of texts without being told which method had been used. 'The feedback consistently showed that readers could not reliably distinguish between them,' according to a company statement. 'In some cases, reviewers even felt the AI-assisted versions were closer in tone and fidelity to the original English manuscript.' However, prominent translators along with a translators' organisation have expressed concern over the initiative. GlobeScribe 'may claim to unlock global access for fiction, but their approach sidelines the very people who make literature resonate across cultures,' said Ian Giles, chair of the Society of Authors' Translators Association. 'Suggesting that AI can match, or even surpass, the nuanced work of human translators on behalf of authors is flat-out wrong.' 'The best literary translations offer more than simple accuracy, more than literal fidelity to the words making up the sentences,' said Polly Barton, writer and translator of works including the bestselling Butter by Asako Yuzuki from Japanese to English. 'They are engaging with the context from which the book has come, and reproducing the pacing, atmosphere, emotional timbre, rhythm, and all the other, less superficially obvious factors that ultimately determine how fulfilling and rich the reading experience is.' Deepa Bhasthi, whose translation of Banu Mushtaq's Heart Lamp from Kannada into English won this year's International Booker prize, said that 'there are many words in Kannada that contain whole cultural worlds in them, where there is as much hidden or implied within a cultural context as is conveyed in literal meaning. And that needs a human being, with an understanding of these visible and invisible worlds, to translate such words'. Asked about GlobeScribe's testing method, Barton said that being a native speaker of a language 'doesn't necessarily equip one to judge all translations from that language with total authority'. Bhasthi added that 'we are not told what kind of texts they were given, what kind of readers the test subjects were'. GlobeScribe's founders said that while they 'recognise that parts of the industry are understandably cautious about what AI might mean for the arts', they 'believe these tools are here to stay and that they should be embraced thoughtfully and responsibly'. They added that AI could enhance creativity and help professional translators 'increase their productivity and output'. The founders 'are clear that this is not about replacing human translators'. Julia Sanches – the translator of works including Boulder by Eva Baltasar from Catalan into English – said: 'Even though I don't think Globescribe can translate the kinds of literary texts I translate, I am gloomy about the emergence of all these new AI 'translation' services. They give the appearance that translation is instant, which devalues my labour, and also that it is mediocre, which could make 'good enough' the new standard for the literary arts. And that's a disservice to both authors and readers.' 'This doesn't just end with translation,' said Barton. 'Maybe translators are at the frontline of people being put out of work by AI technologies, but soon there will be more and more jobs threatened to be wiped out in this way. It lies in our hands to decide whether or not we want this to happen.'
Yahoo
22-05-2025
- Business
- Yahoo
Timekettle Charts Path from L3 to L4 in Real-Time Translation
HONG KONG, May 22, 2025 /PRNewswire/ -- Timekettle, a global leader in AI-powered translation technology, exhibits at the 11th Asia-Pacific Translation and Interpreting Forum (APTIF 11), held in Hong Kong under the theme "Culture, Connectivity and Technology: Translating Communities, Transforming Perspectives." Wilber Wu, Head of Marketing at Timekettle, delivered a speech titled "The Journey of Timekettle: Breaking Language Barriers with Technology", presenting the company's vision and innovation roadmap in real-time AI translation at the event's sub-forum. From Ambition to Innovation Founded in 2016 in Shenzhen, Timekettle has grown from a startup into a global player in AI translation solutions. With operating centers in both China and the United States, the company pioneered Babel OS — the world's first simultaneous AI interpretation system — and developed award-winning hardwares like the W4 Pro AI Interpreter Earbuds and the X1 AI Interpreter Hub. Timekettle's products have served more than 1.5 million users in over 170 countries. Yet as Wu emphasized, Timekettle's true identity lies not in hardware, but in its relentless pursuit of technological evolution — to make real-time multilingual communication as natural and seamless as speaking one's native tongue. The Challenge in Real-Time Translation Wu acknowledged that language remains one of the most persistent barriers in global communication. "Real-time translation is deceptively complex," he noted. "It's not just about translating words, but delivering speech fast, fluently, and with cultural awareness — all without disrupting natural conversation." In use cases such as international business, multi-language classrooms, or humanitarian aid, even a few seconds of delay or a mistranslated sentence can render the experience ineffective. To solve this, Timekettle has been working on improving three critical dimensions: stability, speed, and accuracy. The L1–L5 Model: A Roadmap Toward Seamless Communication In his speech, Wu introduced a technological evolution framework of real-time translation, inspired by autonomous driving standards. The model outlines Timekettle's vision for the technological progress, not just for the company but for the industry as a whole. L1 – Basic Assistance: Word/phrase translation via rule-based or statistical machine translation. Requires manual input and lacks context. L2 – Context-Aware Translation (human-mimicking): Full-sentence translation with neural machine translation (NMT). Still largely manual and non-conversational. L3 – Real-Time Bi-Directional Translation (near-human): Achieved by Timekettle's W4 Pro, offering bidirectional real-time communication with sub-6-second latency, powered by ASR + NMT + TTS. L4 – High-Accuracy Simultaneous Translation (human-equivalent): This next stage will expect delay under 3 seconds, with 98%+ accuracy, and emotional/cultural nuance using multimodal AI and adaptive LLMs. L5 – Fully Autonomous Translation (beyond human): A near-sentient level of AI that matches or surpasses human interpreters — understanding subtext, emotion, dialects, and delivering zero-latency results. Wu likened this evolution to Timekettle's pursuit of a "Babel Fish" — the mythical translator from science fiction — and noted that the company's Babel OS marks the industry's first leap from L3 toward L4, from functional to human-like translation. Looking Ahead: Multimodal AI and AGI Looking to the future, Wu identified multimodal AI and Artificial General Intelligence (AGI) as pivotal to reaching L5. "To truly erase language boundaries, machines must do more than decode words. They must understand gestures, emotions, and cultural subtext — to become communication assistants, not just tools." He outlined Timekettle's current research directions: improving voice recognition in noisy environments, making breakthroughs in text data of certain languages, and enabling machines to better understand cultural nuances and implied meanings in conversations. "Imagine a world," he said, "where translation is so seamless that it becomes invisible; where communication across languages feels as natural as speaking your own. That's the future we're building toward." Bridging Academia and Industry Wu concluded by thanking APTIF for fostering dialogue between the academia and the industry. "Translating technology is not just about algorithms or devices — it's about enabling human connection. As we move from L3 to L4 and beyond, we hope to see researchers and technology professionals join this venture together to make real-time, universal communication a reality." Timekettle's presence at APTIF 11 marks a significant step in aligning cutting-edge AI with the long-standing mission of the translation community - building bridges across cultures through better communication. About Timekettle Since its inception in 2016, Timekettle has been at the forefront of cross-language communication innovation. With award-winning products and a global user base of over one million, Timekettle continues to set new standards in AI translation technology, striving to achieve universal communication freedom. View original content to download multimedia: SOURCE Timekettle Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
CBC
16-05-2025
- Politics
- CBC
The Sunday Magazine for May 18, 2025
This week on The Sunday Magazine with Piya Chattopadhyay: What Trump's Middle East tour means for U.S. foreign policy The White House is championing a diplomatic breakthrough with Syria and deepened investment ties in the Gulf region, following U.S. President Donald Trump's state visits to Saudi Arabia, Qatar and the United Arab Emirates this past week. The New York Times White House correspondent Luke Broadwater, who was on the trip, joins Chattopadhyay to break down the visit's implications for U.S. foreign policy. Then, The Economist's Middle East correspondent Gregg Carlstrom explains the broader impact of Trump's visit in the region, especially as Israel's war with Hamas continues to rage in Gaza, and relations between Israeli Prime Minister Benjamin Netanyahu and Trump reportedly grow sour. Will we still need to learn foreign languages if AI can do the translating? If you're planning a summer vacation abroad and your packing list includes a phrasebook or bilingual dictionary… this might be the last time you need one. Several big tech companies have live translation devices either in the works or already on the market. But tech watchers say they raise larger questions about how AI tools are changing language learning and translation – and what might get lost in the process. WIRED's senior business reporter Louise Matsakis and Muhammad Abdul-Mageed, the Canada Research Chair in Natural Language Processing and Machine Learning at the University of British Columbia, join Chattopadhyay to spell it all out. After a week of clashes over talks, is peace any closer in the Russia-Ukraine War? This past week, Ukrainian President Volodymyr Zelenskyy invited Russian President Vladimir Putin to join him for face-to-face peace talks in Turkey. The Russian leader didn't show. But delegations from both countries did hold their first direct talks in three years, negotiating a prisoner-of-war exchange. Michael Bociurkiw, a global affairs analyst and senior fellow at the Atlantic Council, joins Chattopadhyay to make sense of these latest efforts at bringing a diplomatic end to the years-long conflict. Uncovering the stories that paved the Trans-Canada Highway For many Canadians, the May long weekend means the official kick-off of the summer travel season. And if you're out on a road trip, there's a good chance you may end up sailing down the Trans-Canada Highway. Mark Richardson wants us all to know the often overlooked stories that paved this roughly seven-and-a-half thousand kilometre road system. The automotive journalist speaks with Chattopadhyay about the history, people and politics that shaped "Canada's Main Street", as he explores in his book The Drive Across Canada. If you've ever innocently called someone's summer home a "cabin" or "camp" when they prefer to call it a "cottage," you'll know how evocative those descriptors can be. As cabin/camp/cottage season kicks off, linguist Sali Tagliamonte walks us through the cultural, geographic and economic factors that have shaped the language we use when describing Canada's summertime escape hatches.
Forbes
16-05-2025
- Forbes
These Tiny Earphones Translate Between 15 Languages Using AI
Acer AI TransBuds Acer Language translation is one of the more compelling uses for AI and machine learning tech, and Acer has built it into a pair of earphones, the Acer AI TransBuds. Well, sort of. The Acer AI TransBuds look much like a fairly ordinary pair of open ear headphones, one of the most recent design trends in wireless pairs, but they are made for AI-powered 'two-way voice translation.' And the idea is only one person needs to wear the Acer AI TransBuds. That doesn't entirely make sense until you hear these earphones also rely on a wirelessly connected phone or tablet, which can display the 'live captioning and transcription' results, perhaps while the wearer gets an audio translation. It's this connected device that presumably does all of the real work of translation — most likely via cloud compute — as a pair of earphones is unlikely to be able to hold the processing and battery power required for full on-device translation for some time still. The Acer AI TransBuds have tiny 50mAh batteries, and include a charging case just like a regular true wireless pair. They support Bluetooth 5.4 and have an ear hook design, which is necessary because the actual bud parts don't have an IEM-style tip or even the classic style of earbud designed to rest behind the ear's tragus. That's deliberate, to stop them from blocking you hearing what the other person is saying, even if you can't understand them. Acer says the AI TransBuds support 15 languages at present. And while these are not specified, they are described as 'major languages spoken across Asia, the Americas, and Europe.' The Acer AI TransBuds are not quite as revolutionary as they may seem from a one-sentence description, but it is refreshing to see a household name take on what is still quite a niche area of tech. They might be compared to the Vasco Translator E1 earbuds or the Timekettle M3 Translation Earbuds. All three pairs operate using the same concept, pairing to a smartphone app in order to provide translation. This also means the real worth of the Acer AI TransBuds is in the part Acer has talked little about so far, the software. Acer also has not yet revealed when the AI TransBuds are planned for a release, or where they will be marketed.
