{"slug": "the-machine-translation-newsletter-august-2026", "title": "🌏 The machine translation newsletter - August 2026", "summary": "Google DeepMind launched SL2T, a sign-language-to-text model that adds American Sign Language (ASL) dictation to Pixel 11's Gboard and Live Transcribe, while OpenAI introduced GPT-Live, a voice model replacing Advanced Voice Mode that enables simultaneous listening and speaking with live translation. The US Congress is debating the Language Access for All Act, which would restore language-access requirements revoked in 2025 and prohibit AI from replacing human translators, and Gartner featured ModelFront in its 2026 Hype Cycle for Natural Language and Unstructured Content Technologies.", "body_md": "Hi folks!\n\nLive translation is showing up everywhere this month — Claude’s voice mode, Threads Live Chats, DeepL in Google Meet, OpenAI’s GPT-Live, and even live commentary at a track and field championship. Sign language joined the mix too, with Google DeepMind’s SL2T landing on Pixel phones.\n\nOur community shared fun projects too. We loved them! One app added High Valyrian. Another lets you read EPUBs with side-by-side translation. A third keeps character names and terms consistent when translating novels into Chinese.\n\nEvent season continues. AMTA, AAMT, and MT Marathon are coming up. There's a new Translation AI event at Tech Week in SF in October. And WMT is in November.\n\nEnjoy — and share!\n\n🗞️ **News**\n\nThe\n\n**US Congress** is debating the, which would restore language-access requirements revoked in 2025 and stop AI from replacing[Language Access for All Act](https://www.reddit.com/r/machinetranslation/comments/1v3fzbt/new_proposed_bill_restores_required_translation/)**human translators**.** Gartner**featuresin its 2026[ModelFront](https://www.reddit.com/r/machinetranslation/comments/1vb1zcc/modelfront_featured_in_gartner_hype_cycle_for/)**Hype Cycle** for Natural Language and Unstructured Content Technologies.\n\n🚀 **Launches**\n\n**Google DeepMind** launches, a sign-language-to-text model that adds[SL2T](https://www.reddit.com/r/machinetranslation/comments/1vo9in8/google_deepmind_launches_sl2t_a/)**ASL** dictation to**Pixel 11**’s** Gboard**and** Live Transcribe**.** OpenAI**launches, a voice model that replaces[GPT-Live](https://www.reddit.com/r/machinetranslation/comments/1vlrf2h/openai_launches_gptlive_model_capable_of/)**Advanced Voice Mode**, listens and talks at the same time, and demoed live translation.** Eurovision Sport**and** CAMB.AI**launch[AI-translated live](https://www.reddit.com/r/machinetranslation/comments/1vlsl73/eurovision_sport_launches_aitranslated/)at the[commentary](https://www.reddit.com/r/machinetranslation/comments/1vlsl73/eurovision_sport_launches_aitranslated/)**European Athletics U18 Championships**.** Anthropic**launches[multilingual](https://www.reddit.com/r/machinetranslation/comments/1vatple/anthropic_expands_multilingual_voice_mode_for/)[support for](https://www.reddit.com/r/machinetranslation/comments/1vatple/anthropic_expands_multilingual_voice_mode_for/)[Claude](https://www.reddit.com/r/machinetranslation/comments/1vatple/anthropic_expands_multilingual_voice_mode_for/)[’s](https://www.reddit.com/r/machinetranslation/comments/1vatple/anthropic_expands_multilingual_voice_mode_for/)on[voice mode](https://www.reddit.com/r/machinetranslation/comments/1vatple/anthropic_expands_multilingual_voice_mode_for/)**Opus** and**Sonnet** models.**Translated** launches, adding training based on human feedback plus[Lara 3](https://www.reddit.com/r/machinetranslation/comments/1vayhl3/translated_is_now_releasing_lara_v3/)**image**,** audio**, and** document**translation.** Instagram**and** Facebook**launch[Reels](https://www.reddit.com/r/machinetranslation/comments/1vlrhln/instagram_and_facebook_expand_reel_translation_to/)[translation](https://www.reddit.com/r/machinetranslation/comments/1vlrhln/instagram_and_facebook_expand_reel_translation_to/)in**18 languages**, with** auto-dubbing**and** AI lip-sync**.** TAUS**launches[version 3 of its](https://www.reddit.com/r/machinetranslation/comments/1vkmuqi/taus_launches_version_3_of_their_offtheshelf/)[EPIC](https://www.reddit.com/r/machinetranslation/comments/1vkmuqi/taus_launches_version_3_of_their_offtheshelf/)[quality-estimation](https://www.reddit.com/r/machinetranslation/comments/1vkmuqi/taus_launches_version_3_of_their_offtheshelf/)[model](https://www.reddit.com/r/machinetranslation/comments/1vkmuqi/taus_launches_version_3_of_their_offtheshelf/).**Threads** launches[translation](https://www.reddit.com/r/machinetranslation/comments/1vlr2fc/threads_launches_translation_support_for_live/)[support](https://www.reddit.com/r/machinetranslation/comments/1vlr2fc/threads_launches_translation_support_for_live/)for**Live Chats**.** Amazon**launches, an AI translation[Kindle Translate](https://www.aboutamazon.com/news/books-and-authors/amazon-kindle-translate-books-authors)**service**, in** French**,** Italian**, and** Portuguese**for self-published authors.** DeepL Voice**launches[real-time](https://www.reddit.com/r/machinetranslation/comments/1v7u569/deepl_voice_is_now_available_for_google_meet/)[translated](https://www.reddit.com/r/machinetranslation/comments/1v7u569/deepl_voice_is_now_available_for_google_meet/)for[captions](https://www.reddit.com/r/machinetranslation/comments/1v7u569/deepl_voice_is_now_available_for_google_meet/)**Google Meet**.** Pangeanic**launches, a[MTQE v2](https://www.reddit.com/r/machinetranslation/comments/1vg12rt/we_just_released_pangeanic_mtqe_v2_on_hugging_face/)**quality-estimation** model, on**Hugging Face**.** Azure Translator API**launches[LLM](https://www.reddit.com/r/machinetranslation/comments/1v4ddip/updates_on_the_azure_translator_api_llm_choice/)[choice,](https://www.reddit.com/r/machinetranslation/comments/1v4ddip/updates_on_the_azure_translator_api_llm_choice/)[adaptive MT](https://www.reddit.com/r/machinetranslation/comments/1v4ddip/updates_on_the_azure_translator_api_llm_choice/)[, and](https://www.reddit.com/r/machinetranslation/comments/1v4ddip/updates_on_the_azure_translator_api_llm_choice/)controls.[tone](https://www.reddit.com/r/machinetranslation/comments/1v4ddip/updates_on_the_azure_translator_api_llm_choice/)\n\n🛠️ **Community launches**\n\n, an open-source pipeline for consistent[Wenyi](https://www.reddit.com/r/machinetranslation/comments/1v38f9a/opensourced_wenyi_an_aiassisted_workflow_for/)**terminology** and**character names** in**Chinese** novel translation., a free, open-source[GenSRT](https://www.reddit.com/r/machinetranslation/comments/1vc0tbq/tool_gensrt_free_opensource_whisper_subtitle_app/)**Whisper** subtitle app for**Windows**, with translation and** burn-in**.** Rhetoriq**launches an app for[AI](https://www.reddit.com/r/machinetranslation/comments/1vfihr9/rhetoriq_app_for_dialects/)[rewriting, translation, and dialect conversion](https://www.reddit.com/r/machinetranslation/comments/1vfihr9/rhetoriq_app_for_dialects/), adding**High Valyrian** for**Game of Thrones** fans., an open-source CLI for translating[XliffAI](https://www.reddit.com/r/machinetranslation/comments/1vjn0zd/xliffai_an_opensource_clilibrary_for_translating/)**XLIFF** 2.x files with LLMs.**GianoReader** adds[side-by-side](https://www.reddit.com/r/machinetranslation/comments/1vkgiqk/oc_gianoreader_updates_you_can_now_read_your/)translation on phones and tablets.[EPUB](https://www.reddit.com/r/machinetranslation/comments/1vkgiqk/oc_gianoreader_updates_you_can_now_read_your/)**Olaverse Lab** launches, a specialized model for restoring[DiacNet-1.1](https://www.reddit.com/r/machinetranslation/comments/1vbzr2f/diacnet11_a_specialized_model_for_restoring/)**diacritics** in 10 languages, fine-tuned from Google’s**ByT5**., a free e-book translator supporting[Libyx](https://www.reddit.com/r/machinetranslation/comments/1v1jr00/translating_ebooksepubs/)**43 languages**, cuts costs with** batch mode**and** prompt caching**, and refines quality with a two-model pass., a browser extension that shows an LLM-powered translation in a popup over selected text, using[FTA](https://www.reddit.com/r/machinetranslation/comments/1voemct/fta_extension/)**Google AI Studio**,** OpenRouter**, or local models., a real-time speech translation app for[Meeting Interpreter](https://www.reddit.com/r/machinetranslation/comments/1vpqaf4/i_built_meeting_interpreter_realtime_speech/)**Windows**., a document translator that lets you pick between[DocTranslating](https://www.reddit.com/r/machinetranslation/comments/1vpr429/a_document_translator_that_allows_you_to_choose/)**DeepL**,** Google**,** Azure**, and** Gemini**per job, with OCR support for** handwritten text**., a bot-free AI meeting assistant, adds live translation in[OLVA](https://www.reddit.com/r/machinetranslation/comments/1vq37p0/we_added_realtime_meeting_translation_to_our/)**87 languages** across**Zoom**,** Google Meet**, and** Microsoft Teams**.\n\n📚 **Research**\n\n**Taia Translations** shares[its experience adding](https://www.reddit.com/r/machinetranslation/comments/1uy41yt/someone_here_asked_why_commercial_llm_translator/)[LLM](https://www.reddit.com/r/machinetranslation/comments/1uy41yt/someone_here_asked_why_commercial_llm_translator/)[tiers](https://www.reddit.com/r/machinetranslation/comments/1uy41yt/someone_here_asked_why_commercial_llm_translator/)to its product, citing unreliable segment handling and a shift to**credit-based pricing**.** WMT26**’s[quality-evaluation](https://www.reddit.com/r/machinetranslation/comments/1v80ff2/wmt26_automated_translation_quality_evaluation/)[shared task](https://www.reddit.com/r/machinetranslation/comments/1v80ff2/wmt26_automated_translation_quality_evaluation/)closes for submissions, covering**error detection**,** score prediction**, and** error-free classification**.A\n\n[technical walkthrough](https://www.reddit.com/r/machinetranslation/comments/1v86rfd/built_trained_a_transformer_from_scratch_in_pure/)of building a**Transformer** from scratch in**PyTorch**, for** English-to-Tamil**translation.\n\n❔ **Questions**\n\nHow do you preserve\n\nwhen translating[table format](https://www.reddit.com/r/machinetranslation/comments/1v2lwf5/traducci%C3%B3n_de_archivos_de_ingl%C3%A9s_a_espa%C3%B1ol/)**English** to**Spanish** files?Is a\n\ntranslation trick real?[Google Lens](https://www.reddit.com/r/machinetranslation/comments/1v33y7y/i_discovered_this_by_accident_is_it_true_i_just/)Which real-time\n\napp works for[voice translation](https://www.reddit.com/r/machinetranslation/comments/1v40fcg/what_is_the_best_realtime_voice_translation_app/)**English** and**Hungarian**?Has\n\ngotten worse?[DeepL](https://www.reddit.com/r/machinetranslation/comments/1vcq5p5/has_deepl_translate_gotten_worse/)What\n\n[AI](https://www.reddit.com/r/machinetranslation/comments/1vd6510/what_are_ai_instructions_for_making_best/)do people use for translation?[instructions](https://www.reddit.com/r/machinetranslation/comments/1vd6510/what_are_ai_instructions_for_making_best/)Can anyone beta-test a\n\ntranslation on a[Hindi](https://www.reddit.com/r/machinetranslation/comments/1vdxm5a/namaste_i_create_a_lord_of_the_rings_translation/)**Lord of the Rings**-themed site?Can anyone help with a\n\ntranslation for a research project?[Sanskrit](https://www.reddit.com/r/machinetranslation/comments/1vfj4af/sanskrit_translation_for_research_project/)What advice do people have on\n\nfor an MT internship project?[preprocessing](https://www.reddit.com/r/machinetranslation/comments/1vg3f1p/help_on_machine_translation/)Which\n\nsubtitle tool should I use?[local STT/ASR](https://www.reddit.com/r/machinetranslation/comments/1vi92k3/subtitle_tool_req_wlocal_stt_asr_issues/)What workflow do people use for translating\n\nwith AI?[long-form books](https://www.reddit.com/r/machinetranslation/comments/1vjata2/what_is_the_best_workflow_for_translating/)What do you learn building an\n\n-based translation setup[LLM](https://www.reddit.com/r/machinetranslation/comments/1uz3py5/lessons_ive_learned_so_far_from_trying_to/)**on a budget**?Which MT tools can translate\n\ninto English?[medieval Tibetan](https://www.reddit.com/r/machinetranslation/comments/1vov7al/best_way_to_translate_from_medieval_tibetan_into/)Does\n\nactually work?[machinetranslation.com](https://www.reddit.com/r/machinetranslation/comments/1vp8m4q/does_machinetranslationcom_actually_work/)Why doesn’t\n\nuse a voice transcriber as accurate as[WhatsApp](https://www.reddit.com/r/machinetranslation/comments/1vrgxfu/i_wish_there_was_a_multilingual_voice/)**Instagram**’s?\n\n⏰️ **Deadlines**\n\ncall for papers — 5 September[MT Marathon 2026](https://www.reddit.com/r/machinetranslation/comments/1sadded/mt_marathon_2026_karlsruhe_germany/)\n\n📆 **Events**\n\n— 31 August - 2 September — Québec City 🇨🇦[AMTA 2026](https://machinetranslate.org/amta-2026)— 13 - 14 September — Online / Nara 🇯🇵[AAMT 2026](https://machinetranslate.org/aamt-2026)— 14 - 18 September — Karlsruhe 🇩🇪[MT Marathon 2026](https://www.reddit.com/r/machinetranslation/comments/1sadded/mt_marathon_2026_karlsruhe_germany/)— 8 October — San Francisco 🇺🇸[Translation AI at Tech Week](https://www.modelfront.com/events/translation-ai-sf-tech-week-2026)— November — Budapest 🇭🇺[WMT 2026](https://machinetranslate.org/wmt26)\n\n💻 **Jobs**\n\n[Head of Global Language Services](https://www.reddit.com/r/machinetranslation/comments/1uz75i9/head_of_global_language_services_lse_at_interpol/)[(LSE)](https://www.reddit.com/r/machinetranslation/comments/1uz75i9/head_of_global_language_services_lse_at_interpol/)at**INTERPOL**— Lyon 🇫🇷[Manager](https://www.reddit.com/r/machinetranslation/comments/1v1ubqa/manager_new_localization_technology_at_deluxe/)[, New Localization Technology](https://www.reddit.com/r/machinetranslation/comments/1v1ubqa/manager_new_localization_technology_at_deluxe/)at**Deluxe**— Burbank 🇺🇸[Machine Learning Manager](https://www.reddit.com/r/machinetranslation/comments/1v3juob/machine_learning_manager_localization_algorithms/)[, Localization Algorithms](https://www.reddit.com/r/machinetranslation/comments/1v3juob/machine_learning_manager_localization_algorithms/)at**Netflix**— Los Gatos 🇺🇸[Program Manager](https://www.reddit.com/r/machinetranslation/comments/1v3jypn/program_manager_localization_testing_at_meta_san/)[, Localization Testing](https://www.reddit.com/r/machinetranslation/comments/1v3jypn/program_manager_localization_testing_at_meta_san/)at**Meta**— San Francisco 🇺🇸at[Localization Program Manager](https://www.reddit.com/r/machinetranslation/comments/1v3jzyz/localization_program_manager_at_meta_menlo_park_ca/)**Meta**— Menlo Park 🇺🇸", "url": "https://wpnews.pro/news/the-machine-translation-newsletter-august-2026", "canonical_source": "https://newsletter.machinetranslate.org/p/the-machine-translation-newsletter-aff", "published_at": "2026-08-19 19:11:38+00:00", "updated_at": "2026-08-19 19:45:04.502930+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "natural-language-processing", "ai-products", "ai-policy"], "entities": ["Google DeepMind", "SL2T", "Pixel 11", "Gboard", "Live Transcribe", "OpenAI", "GPT-Live", "US Congress"], "alternates": {"html": "https://wpnews.pro/news/the-machine-translation-newsletter-august-2026", "markdown": "https://wpnews.pro/news/the-machine-translation-newsletter-august-2026.md", "text": "https://wpnews.pro/news/the-machine-translation-newsletter-august-2026.txt", "jsonld": "https://wpnews.pro/news/the-machine-translation-newsletter-august-2026.jsonld"}}