Atlas AI Institute develops open, public-interest digital infrastructure — from AI policy observatories and civic technology tools to citizen-facing services — enabling evidence-based governance across Bangladesh, South Asia, and the Global South.
Open-source, AI-powered tools built by Atlas AI Institute — from policy auditing and governance monitoring to financial-crime detection and disaster early warning. Priority-ranked and publicly available.
A 97-nation open AI policy repository for the Global South — EU AI Act, UNESCO Ethics, OECD Principles, G20 Framework, and ITU standards, locally contextualised per governance maturity and cultural baseline.
A comparative AI policy benchmarking and scoring platform ranking national governance frameworks side by side — built for policymakers, researchers, and journalists tracking AI regulation across the Global South and beyond.
A 9-dimension AI-powered policy audit engine covering Cultural Alignment, Economic Feasibility, Human Rights, Corruption Risk, SDG Alignment, Institutional Capacity, Gender Equity, Environmental Impact, and Political Feasibility. Powered by World Bank WGI, TI CPI 2023, J-PAL MIT, and the Hofstede 6D model.
Enterprise anti-money laundering detection using NetworkX graph analysis and Claude AI forensic narratives to generate SAR/STR reports. Detects mule clusters, bot smurfing, shell company layering, PEP exposure, and crypto offramp evasion. FATF aligned.
Predictive Anticipatory Governance for floods, cyclones, and droughts — Apache Kafka streaming, BMD + BWDB APIs, IoT gauges, Isolation Forest anomaly detection, and bilingual Bangla–English LLM alerts on a ~50KB SVG dashboard.
Offline-first, AES-256 encrypted beneficiary data collection for field operations — USB key-gated access, encrypted backups, immutable audit logs, GDPR Art. 9 compliance, and alignment with the ICRC Data Protection Framework. Zero internet required.
A diagnostic scorecard for assessing a country's or institution's readiness to adopt, regulate, and govern AI — evaluating infrastructure, talent, data ecosystems, regulatory maturity, and ethics capacity against international benchmarks.
All repositories are publicly available on GitHub →
Publicly deployed civic initiatives from Atlas AI Institute — citizen services, misinformation defense, and national AI policy infrastructure for Bangladesh.
An AI-powered civic assistant helping citizens navigate government and public services across Bangladesh — from finding the right office and required documents to understanding procedures, in plain language.
Bangladesh's first AI-powered fact-checking platform — multi-agent misinformation detection using CrewAI orchestration, Bengali NLP, and live RSS ingestion. Built on a 773K-entry validated, labeled dataset. Bengali + English, IFCN aligned.
A comprehensive repository of governance frameworks supporting the responsible, secure, and trustworthy development and deployment of AI in Bangladesh, South Asia, and the Global South — covering policy, regulation, safety, and transparency.
Peer-reviewed and open-access publications underpinning our tools and projects — available on Zenodo, ResearchGate, and atlasai.institute.
Benchmarking policy maturity, institutional readiness, and regulatory design across the world's most advanced AI ecosystems. Open-access, peer-reviewed, published on Zenodo and ResearchGate.
A legally precise, modular reference architecture for digital sovereignty and technological self-determination — built for low- and middle-income nations to assert data sovereignty and reject passive technology dependency.
Atlas AI Institute is founded and led by Nafiul Ahmad Rafi →
