📣NEW! SmolTalk-Chinese: high-quality multi-task dataset for Chinese conversational AI, covering 19 task types: advice-seeking, code gen, daily chat! Dialogue model training & LLM fine-tuning 🧡 📎 Available at datacollective.mozillafoundation.org/datasets/cmk... ⚠️ Academic/non-commercial use only

📣🌍 Meet the Ewondo_Fong_ALCAM-MultimodalDataset! Highlighting Fong; a variant of Ewondo speech that’s widely used but rarely documented. Native-speaker audio, example sentences + Fong–French text to support language learning and speech tech 🧡 🔗 datacollective.mozillafoundation.org/datasets/cmk...

Thinking back to how many times we picked up a new information tool (and often put down an old one)… Also, announcing EM Lewis-Jong from @mozdatacollective.bsky.social as a speaker for #PfP✨London

📣 New dataset! 🇫🇷 Large-scale collection of 400,000+ real user conversations generated on the compar:IA platform. #opensource and proprietary systems developed within the French Ministry of Culture. 👏 datacollective.mozillafoundation.org/datasets/cmk... @info.gouv.fr @mozilla.org

Who gets to shape the tech we use? Since launching in Nov datacollective.mozillafoundation.org has been so excited about the educators, media orgs and technologists joining us! New in: • Teacher-led bias dataset (Argentina) • 70 hours of Luhya speech data (Kenya) • Radio archives (Northern Uganda)

Communities built the world’s largest open voice dataset. Now they’re building a new model for AI data — one rooted in choice, transparency, and collective power. Mozilla Data Collective is live🎉 ✳️ Join the community mzl.la/44aTEoe @mozdatacollective.bsky.social #MozillaDataCollective

Will you be in Melbs for @pyconau.bsky.social #PyConAU #PyConAU2025 in the next few days? Don't miss MDC's @kathyreid.au speaking on Sun 14th in Ballroom 1 at 1410hrs - introducing a better way to provide and source #data for #AI. Sneak peek: community.mozilladatacollective.com/your-dataset...

Hello world! We're the Mozilla Data Collective - and we want to rebuild the #AI #data #ecosystem - with #communities at the centre. 🚫 Unconsented web scraping 🚫 Sign license deal with megacorp ✅ The Mozilla Data Collective Launching Wednesday! 🚀 💫 datacollective.mozillafoundation.org