Building Resposible AI for Social sector - Kaapi & Tattle

​Building Resposible AI for Social sector: Introducing Open-Source AI Guardrails

​AI assistants and chatbots are becoming increasingly common across the social sector, helping organizations engage beneficiaries, improve service delivery, and scale support. But as adoption grows, so do concerns around privacy, harmful content, inaccurate responses, and AI hallucinations.

​To address these challenges, Project Tech4Dev and Tattle are building open-source AI guardrails designed specifically for the Indian social sector. These guardrails help organizations build AI applications that are safer, more reliable, and better suited for beneficiary-facing use cases.

​In this webinar, we'll introduce the solution, demonstrate how it works, and explore how nonprofits can use, test, and contribute to its development.

​What we'll cover

• ​Why AI guardrails matter for nonprofits

• ​Research and design considerations for Indian languages

• ​Live demonstrations across healthcare and education use cases

• ​Opportunities to pilot, co-build, and strengthen the solution

​We'll also hear from The Apprentice Project (TAP), who will share their experience implementing AI guardrails, including their use case, the experiments they conducted, and key learnings from the process.