Tilde’s Post

In a world of over 7,000 languages, AI's fluency in only 100 highlights a gap in our digital future. This linguistic divide isn't just about words - it's about access, opportunity, and the power to participate in an increasingly AI-driven global economy. The implications of this gap are profound. In Africa, where a third of the world's youth will reside by 2050, not a single one of the top 34 internet languages originates. This disparity threatens to exclude millions from the digital revolution, potentially deepening global inequalities rather than bridging them. The challenge before us is not just technological but deeply human - how do we make sure that the AI systems shaping our world can understand and serve all of humanity? Ensuring AI's linguistic diversity isn't merely a technical challenge—it's a gateway to equitable progress. By teaching our digital assistants to speak the languages of all communities, we're not just improving AI; we're democratising the future, one word at a time. https://lnkd.in/eE_Y4ZRN #AI #Linguistics #Language

The Aya project by Cohere for AI is making significant strides in expanding access to foundation models in global languages beyond English. The initiative, launched by Cohere's research arm, introduced the Aya 101 large language model covering 101 languages with a 13-billion-parameter model. Additionally, the release of the Aya dataset aims to facilitate access to other languages for model training. Aya Expanse, built on the success of Aya 101, reflects a sustained effort to enhance how AI caters to languages worldwide by reimagining fundamental machine learning components. Cohere expressed that the enhancements in Aya Expanse stem from an unwavering commitment to bridging the language gap. Key breakthroughs such as data arbitrage, preference training for improved performance and safety, and model merging have been pivotal in shaping the project's trajectory. https://lnkd.in/gjHCSJpR #AI #LanguageModels #Cohere #AyaProject #MachineLearning

Still thinking about Michael Running Wolf and his talk from yesterday's O'Reilly #GenAI Superstream. Learn more about the work he's doing at http://firstlanguages.ai. Support the preservation of Indigenous languages and the work they're doing to build #AI tools based on those languages.

This is quite commendable especially because “there are thousands of languages in the world, 1,000 to 2,000 of them in Africa alone: it’s estimated that the continent accounts for one-third of the world’s languages.” #GenAI #legalinnovation #legaltech #lawyer

Cohere has just launched Aya Expanse 8B and 35B, two powerful multilingual AI models as part of the Aya project, now available on Hugging Face! 🚀 These models are designed to bridge the global language gap in AI, supporting 23 languages with groundbreaking advancements. The Aya Expanse models outshine competitors, including Google, Mistral AI, and Meta, in multilingual benchmarks. Cohere’s research in data arbitrage and preference training ensures these models aren’t just powerful but culturally tuned for safe, globally inclusive AI. #aipartnershipscorp #artificialintelligence #machinelearning #language #modlels #enterprise #llm #genai

Google Expands Gemini’s In-Depth Research Mode to 40 Languages: A Deep Dive into the Future of Multilingual AI In a world where digital communication transcends geographic and linguistic barriers, Google is taking a significant leap forward with the...

📝 Introducing our latest report – “AI speaks Polish. The ecosystem of open language models in Poland”! It presents research and conclusions of Alek Tarkowski (Open Future Foundation), Kuba Piwowar (Fundacja Centrum Cyfrowe), and Michał Owczarek (Uniwersytet SWPS). The goal of the report is to provide a case study of Poland’s ecosystem for creating open AI models for the Polish language 🇵🇱 Small language models are filling the gap left by large commercial models, which are not adapted to the Polish language or cultural nuances. The work on these models serves as an example of effectively creating alternatives to dominant entities. The report focuses on two key projects: building the SpeakLeash | Spichlerz language corpus and using it to create the Bielik model, as well as the activities of the #PLLuM consortium (Polish Large Language Model). Based on interviews with the creators of Polish models, the authors outline the development processes and the challenges they presented, and summarise the lessons learned from the achievements so far. 👉 See the full report on our website – in Polish now, and in English next week! 🔗 https://lnkd.in/d7JTr4G8 __ Image: Portrait of Adam Mickiewicz, Austrian National Library [Public Domain, via Europeana.eu]

Orange is partnering with OpenAI and Meta to improve AI large language models (LLMs) for regional languages in Africa that are currently unsupported by GenAI models in the Middle East and Africa. Also Read: OpenAI Partners Global Media Giant for...

In a remarkable stride towards inclusivity in AI, the release of Cohere for AI’s Aya Expanse models marks a pivotal moment for multilingual language processing. Aya Expanse, with its 8B and 32B models, seeks to narrow the linguistic divide by providing accessible and adaptable AI tools for underserved languages. With open weights, these models empower researchers and developers around the globe to innovate without limitations. The significance of this development extends beyond technology; it's a step towards democratizing AI, ensuring all linguistic communities can share in its advancements. Aya Expanse's ability to work across diverse languages like Swahili and Bengali is transformative, outperforming peers in low-resource language benchmarks by a substantial margin. This initiative underscores the importance of accessibility and collective effort in technology. It challenges us to consider: How can we further support language diversity in AI to foster a truly inclusive digital future? Your insights and experiences could contribute to this vital conversation. Original article: [Cohere for AI Releases Aya Expanse (8B & 32B)](https://lnkd.in/dQrCRteJ

I recommend reading the report on the development of genAI in Poland. I was involved in writing it with Kuba Piwowar and Alek Tarkowski, we talked to leaders and experts in the field. My favorite conclusion is why it is worthwhile to build Polish LLMs: because the process creates know-how in the industry and allows institutions to store and process their data locally.