evidoria

← Back to browse

Good practice Imported

aprendIA — An AI Teaching Assistant for Climate-Disaster-Affected Coastal Bangladesh

Bangladesh · Dhaka (deployed in climate-vulnerable coastal districts) · See the Bangladesh profile

Evidence: Descriptive / self-reported Top 73% 40/100 · Ask Evidence Copilot about this practice

OpenAI-funded IRC chatbot 'aprendIA' gives teachers in Bangladesh's climate-vulnerable coastal districts lesson plans and disaster-risk content via WhatsApp; built in under 4 weeks, early testing found 78% of educators rated it intuitive and useful, engaging 3-4 times weekly.

250,000 USD
OpenAI funding to IRC for aprendIA (3-country program) (May 2024)
78 %
Educators rating the platform intuitive and easy to use
78 %
Educators believing it would help manage their classrooms
3-4 sessions/week, 30-40 min each
Weekly engagement frequency reported by educators
6.7 million incl. 3 million children
People needing humanitarian assistance in Bangladesh's coastal districts (2024) (2024)
aprendIA — An AI Teaching Assistant for Climate-Disaster-Affected Coastal Bangladesh

Details

Maturity
Pilot
Promoter
International Rescue Committee (IRC) Airbel Impact Lab, funded by OpenAI
Period
May 2024 – present (2026)
Keywords
generative AI, teacher support, climate-resilient education, crisis and emergency education, early childhood education

Context

In May 2024, OpenAI awarded the International Rescue Committee (IRC) $250,000 to build aprendIA, a generative-AI chatbot (built on ChatGPT and delivered via low-tech messaging such as WhatsApp) supporting teachers and caregivers, rather than tutoring students directly, in crisis- and climate-affected settings in Bangladesh, Nigeria and Colombia. In Bangladesh, IRC's Airbel Impact Lab designed and launched an early-childhood-education version in under four weeks, targeting educators in climate-vulnerable coastal districts, where an estimated 6.7 million people needed humanitarian assistance in 2024, including 3 million children.

Objectives

Give teachers and caregivers lesson plans, disaster-risk-reduction content and classroom-management support, rather than tutoring students directly.

Activities

IRC's Airbel Impact Lab designed and launched the tool in under four weeks; teachers are reported to review the AI tutor's chat logs to adjust lesson plans to individual students' needs.

Results

In IRC's own user testing, participating educators engaged with the chatbot 3-4 times a week for 30-40 minutes per session and completed 2-3 courses each; 78% rated the platform intuitive and easy to use, and 78% believed it would help them manage their classrooms. IRC was named to Fast Company's World's Most Innovative Companies of 2025 in part for this work.

Conclusions

IRC and OpenAI's stated ambition — reaching at least 10,000 teachers/caregivers and 500,000 students across all three countries — is a target, not a confirmed result, and no Bangladesh-specific reach figures have been published. This should be read as a promising, rapidly-deployed humanitarian tool with encouraging early user feedback, not yet as a proven-at-scale intervention.

Implementation
Indicative cost
Low (< €50k) — $250,000 total OpenAI grant to IRC covering the tool's development for Bangladesh, Nigeria and Colombia combined; no Bangladesh-specific cost breakdown has been published.
Time to results
Short (< 1 year) — Built and launched in under four weeks starting May 2024; ongoing as of 2026 reporting.
Staffing & skills
IRC Airbel Impact Lab (design and deployment team), Participating teachers/educators (review chatbot logs to adjust lesson plans)

Conditions for success

  • Low-tech delivery via WhatsApp messaging to work in low-connectivity, climate-vulnerable areas
  • Built directly for teacher/caregiver support rather than direct student tutoring

Common failure modes

  • No Bangladesh-specific reach figures (teachers/students actually using the tool) have been published
  • Stated 3-country targets (10,000 teachers/caregivers, 500,000 students) are unconfirmed goals, not achieved results

Commonly funded by

Philanthropic / foundation funding

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful