Open Knowledge Foundation Nepal

29/07/2026

If you've been curious about AI safety, alignment, or governance, this is worth your time. AI is already shaping how we work, communicate, and make decisions, often before we notice it. The team at AI Safety Nepal has put together the country's first AI safety fellowship, 𝗘𝗽𝗼𝗰𝗵 𝗭𝗲𝗿𝗼: 𝗔𝗜 𝗦𝗮𝗳𝗲𝘁𝘆 𝗙𝗲𝗹𝗹𝗼𝘄𝘀𝗵𝗶𝗽, and applications are open now and closing soon.

Over 5 weeks, you'll discuss how AI systems fail, strategies to make them go well, and who gets to decide what's deployed. No technical background needed. Open to anyone: researchers, engineers, lawyers, policymakers, students, or curious minds from any field.

🖇️ 𝗔𝗽𝗽𝗹𝘆 𝘁𝗵𝗿𝗼𝘂𝗴𝗵: https://airtable.com/appD3ohg0gdE3IPz4/pagUoOCii6OPMWPZd/form
⏱️𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝗰𝗹𝗼𝘀𝗲 𝗔𝘂𝗴𝘂𝘀𝘁 𝟭, 𝟮𝟬𝟮𝟲, 𝟭𝟭:𝟱𝟵 𝗣𝗠 𝗡𝗣𝗧.

📢The Epoch Zero: AI Safety Nepal fellowship is now open for applications!

We’re launching the first ever AI safety fellowship in Nepal. Join us to understand the risks and opportunities of advanced AI. Together, we read core texts, examine key debates, and explore the question: how do we ensure increasingly capable AI systems remain aligned with human values?

What you get:

→ 5 weeks self-paced reading on AI safety
→ Weekly discussion-based learning session, facilitated by researchers & engineers working in AI safety, and with participants from different disciplines
→ Guest lectures from AI safety researchers and governance professionals
→ Interested participants can go on to collaborate with peers on a final project: technical research or policy/governance

And who should apply?

We welcome participants from diverse backgrounds. Whether you are in CS, law, psychology, governance, philosophy, economics, or any other field, and whether you are a student or professional, we are looking for interdisciplinary perspectives.

🖇️Apply now: https://airtable.com/appD3ohg0gdE3IPz4/pagUoOCii6OPMWPZd/form
📅Apply by: August 01, 2026 11:59PM NPT

27/07/2026

At the Digital Democracy Summit 2026, Open Knowledge Foundation Nepal co-led a session with Speak Up Nepal on 𝐂𝐢𝐯𝐢𝐜 𝐓𝐞𝐜𝐡, 𝐀𝐈 𝐚𝐧𝐝 𝐀𝐜𝐜𝐨𝐮𝐧𝐭𝐚𝐛𝐢𝐥𝐢𝐭𝐲. Rather than a panel, we ran a hands-on mapping exercise with government, civil society, private sector, and media participants 🗺️

The session surfaced where the opportunities are, where the gaps remain, and the tools participants want to see built.

🔗 𝐑𝐞𝐚𝐝 𝐭𝐡𝐞 𝐟𝐮𝐥𝐥 𝐟𝐢𝐧𝐝𝐢𝐧𝐠𝐬: https://oknp.org/blogs/mapping-nepals-civic-tech-and-ai-landscape-key-takeaways-from-the-digital-democracy-summit-2026

02/07/2026

🚨 𝐍𝐞𝐰 𝐑𝐞𝐩𝐨𝐫𝐭: 𝐖𝐡𝐞𝐧 𝐀𝐥𝐠𝐨𝐫𝐢𝐭𝐡𝐦𝐬 𝐃𝐞𝐜𝐢𝐝𝐞: 𝐅𝐢𝐥𝐭𝐞𝐫𝐞𝐝 𝐕𝐨𝐢𝐜𝐞𝐬 𝐟𝐫𝐨𝐦 𝐍𝐞𝐩𝐚𝐥

How do automated content moderation systems shape what Nepali creators, journalists, and activists can say online? Our new diagnostic report digs into platform transparency data, community voices, and where the transparency, language, and accountability gaps lie.

📄 𝐑𝐞𝐚𝐝 𝐭𝐡𝐞 𝐟𝐮𝐥𝐥 𝐫𝐞𝐩𝐨𝐫𝐭: https://oknp.org/resources/when-algorithms-decide-diagnostic-report

22/06/2026

When a platform removes your content, suppresses your reach, or cuts your income overnight, and gives you no explanation and no one to call, what do you do?

This is the reality for creators, journalists, and activists navigating algorithmic moderation in Nepal. Our new blog post draws on the findings of "𝐖𝐡𝐞𝐧 𝐀𝐥𝐠𝐨𝐫𝐢𝐭𝐡𝐦 𝐃𝐞𝐜𝐢𝐝𝐞𝐬: 𝐅𝐢𝐥𝐭𝐞𝐫𝐞𝐝 𝐕𝐨𝐢𝐜𝐞𝐬 𝐟𝐫𝐨𝐦 𝐍𝐞𝐩𝐚𝐥," documenting the transparency, language, and accountability gaps that affect millions of Nepali voices every year.

The numbers are stark. The stories behind them are starker. 👇

𝐑𝐞𝐚𝐝 𝐭𝐡𝐞 𝐟𝐮𝐥𝐥 𝐩𝐨𝐬𝐭: https://oknp.org/blogs/when-algorithm-decides-what-nepals-digital-creators-are-up-against

19/06/2026

"𝐃𝐨 𝐲𝐨𝐮 𝐡𝐚𝐯𝐞 𝐩𝐫𝐨𝐨𝐟?" 🔍

She reported the harassment. The evidence was gone. She was told to come back with proof.

This happens because platforms have no structured dataset of 𝐍𝐞𝐩𝐚𝐥𝐢 𝐡𝐚𝐫𝐚𝐬𝐬𝐦𝐞𝐧𝐭 𝐥𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐭𝐨 𝐦𝐨𝐝𝐞𝐫𝐚𝐭𝐞 𝐚𝐠𝐚𝐢𝐧𝐬𝐭. 🚫🗂️

MuLCAM is an open linguistic corpus for Technology-Facilitated Gender-Based Violence (TFGBV), covering Nepali, Romanized Nepali, and mother tongues across 12 harm categories. The annotated datasets are freely available for researchers, developers, and platform safety teams. 📂🌐

We are building moderation tools on top of the corpus and looking for collaborators. Write to [email protected]. 🤝

𝐌𝐨𝐫𝐞 𝐝𝐞𝐭𝐚𝐢𝐥𝐬: https://oknp.org/blogs/introducing-mulcam-building-the-language-foundation-for-nepals-digital-safety-needs

18/06/2026

𝐂𝐨𝐧𝐭𝐫𝐢𝐛𝐮𝐭𝐢𝐧𝐠 𝐭𝐨 𝐚𝐧𝐝 𝐞𝐱𝐩𝐥𝐨𝐫𝐢𝐧𝐠 𝐍𝐞𝐩𝐚𝐥'𝐬 𝐨𝐩𝐞𝐧 𝐝𝐚𝐭𝐚 𝐣𝐮𝐬𝐭 𝐠𝐨𝐭 𝐞𝐚𝐬𝐢𝐞𝐫. 🇳🇵

We shipped two new upgrades to Open Data Nepal:
📤 𝐒𝐞𝐚𝐦𝐥𝐞𝐬𝐬 𝐏𝐮𝐛𝐥𝐢𝐬𝐡𝐢𝐧𝐠: Sign in with Google, GitHub, or email and publish your dataset in minutes. The AI assistant is also available in the dashboard if you need guidance along the way.
📊 𝐂𝐨𝐦𝐩𝐫𝐞𝐡𝐞𝐧𝐬𝐢𝐯𝐞 𝐃𝐚𝐭𝐚 𝐄𝐱𝐩𝐥𝐨𝐫𝐞𝐫: View any dataset visually, build your own charts with a drag-and-drop interface, and explore trends without downloading a single file.

Easier to contribute. Easier to explore. Open to everyone. 💡

👉 𝐑𝐞𝐚𝐝 𝐭𝐡𝐞 𝐟𝐮𝐥𝐥 𝐛𝐥𝐨𝐠: https://oknp.org/blogs/publishing-and-exploring-data-just-got-easier-on-open-data-nepal

17/06/2026

One of the biggest barriers to using open data is not access. It is usability. 🔓

Today, we are taking a step toward changing that. 🇳🇵

𝐀𝐈 𝐢𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐢𝐨𝐧 𝐢𝐬 𝐧𝐨𝐰 𝐥𝐢𝐯𝐞 𝐨𝐧 𝐎𝐩𝐞𝐧 𝐃𝐚𝐭𝐚 𝐍𝐞𝐩𝐚𝐥. 🚀 Anyone can now visit the portal and ask questions about Nepal's public datasets in plain English and get real, contextual answers back.

🔍 Data Discovery
📖 Data Definition
📊 Data Insights

All in one place. No technical background needed.

👉 𝐄𝐱𝐩𝐥𝐨𝐫𝐞 𝐢𝐭: https://opendatanepal.com
📝 𝐑𝐞𝐚𝐝 𝐭𝐡𝐞 𝐟𝐮𝐥𝐥 𝐛𝐥𝐨𝐠: https://oknp.org/blogs/stop-searching-start-asking-ai-integration-is-now-live-on-open-data-nepal

03/06/2026

Thank you to everyone who joined us at the 𝐍𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐑𝐨𝐮𝐧𝐝𝐭𝐚𝐛𝐥𝐞 𝐨𝐧 𝐀𝐥𝐠𝐨𝐫𝐢𝐭𝐡𝐦𝐬 𝐚𝐧𝐝 𝐕𝐨𝐢𝐜𝐞𝐬 on May 31 at Shangrila Blu Hotel, Lalitpur.

From research findings to open discussion, the conversations in the room reminded us why this work matters. 𝐘𝐨𝐮𝐫 𝐯𝐨𝐢𝐜𝐞𝐬, 𝐩𝐞𝐫𝐬𝐩𝐞𝐜𝐭𝐢𝐯𝐞𝐬, 𝐚𝐧𝐝 𝐜𝐨𝐦𝐦𝐢𝐭𝐦𝐞𝐧𝐭𝐬 𝐭𝐨 𝐩𝐥𝐚𝐭𝐟𝐨𝐫𝐦 𝐚𝐜𝐜𝐨𝐮𝐧𝐭𝐚𝐛𝐢𝐥𝐢𝐭𝐲 mean a lot to us.

We look forward to the next steps together.

cc Pulitzer Center

27/05/2026

𝐀𝐥𝐠𝐨𝐫𝐢𝐭𝐡𝐦𝐢𝐜 𝐬𝐢𝐥𝐞𝐧𝐜𝐢𝐧𝐠 𝐢𝐬𝐧'𝐭 𝐫𝐚𝐧𝐝𝐨𝐦. 𝐈𝐭 𝐟𝐨𝐥𝐥𝐨𝐰𝐬 𝐚 𝐩𝐚𝐭𝐭𝐞𝐫𝐧.

Women, activists, journalists, and linguistic minorities are disproportionately affected — not by coincidence, but because they were underrepresented in the data and teams that built these systems.

In Nepal, with 120+ languages and enormous regional diversity, this isn't an edge case. 𝐈𝐭 𝐝𝐞𝐬𝐜𝐫𝐢𝐛𝐞𝐬 𝐚 𝐬𝐢𝐠𝐧𝐢𝐟𝐢𝐜𝐚𝐧𝐭 𝐩𝐨𝐫𝐭𝐢𝐨𝐧 𝐨𝐟 𝐭𝐡𝐞 𝐩𝐞𝐨𝐩𝐥𝐞 𝐰𝐡𝐨𝐬𝐞 𝐯𝐨𝐢𝐜𝐞𝐬 𝐦𝐚𝐭𝐭𝐞𝐫 𝐦𝐨𝐬𝐭 𝐢𝐧 𝐩𝐮𝐛𝐥𝐢𝐜 𝐝𝐢𝐬𝐜𝐨𝐮𝐫𝐬𝐞.

Open Knowledge Nepal is documenting these patterns — so they can be named, understood, and challenged. 🔍