AI Safety Zambia / Est. 2026

MAKING AI
SAFE FOR
ZAMBIA.

We red team frontier AI models, run a public harm observatory where Zambians report AI-driven scams and deepfakes, and build governance frameworks that translate policy into real-world protection.

Why We Exist

Zambia is building with AI.
But who's building it safely?

AI is transforming how Zambians bank, communicate, and access services. But the safety infrastructure to protect people from AI-powered threats does not exist yet.

We are building it before the gap becomes a crisis.

No Safety Infrastructure

The national strategy prioritizes AI growth but leaves red teaming, harm reporting, and safety research entirely absent. Communities are exposed.

AI Threats Are Already Here

WhatsApp voice-clone scams, deepfake political content, and AI-driven mobile money fraud impact Zambians daily with no formal mechanism to report or combat them.

Global Labs, Local Blind Spots

Frontier AI systems are built without Zambian context, missing local languages, cultural nuance, and awareness of region-specific exploitation vectors.

Strategy Without Execution

National AI policy exists on paper, but without technical safety expertise on the ground, governance frameworks remain theoretical. We bring the execution layer.

What We Do

As AI transforms Zambia,
no one gets left unprotected.

01

AI Harm Observatory

A public platform where Zambians report AI-driven scams, deepfakes, and discrimination. Every report feeds our safety research pipeline and policy recommendations.

Free for the Public
02

Red Teaming & Audits

We stress-test AI systems deployed in Zambia. Our founders bring direct experience from evaluating GPT-4o, Deep Research, and Claude at frontier labs.

Labs & Enterprises
03

Policy & Governance

Bridging the National AI Strategy and African Union frameworks with operational policy. We translate research into legislation and enterprise standards.

Government & Ministries
04

Capacity Building

Workshops, executive briefings, and open research publications building the next generation of African AI safety researchers and informed leaders.

Institutions & Universities

Need expert AI safety capability for your organization or government?

Start a Conversation
Threat Intelligence

The threats are
already here.

These are real incidents affecting Zambians today. AI-powered scams, deepfakes, and fraud are not theoretical. They are active, growing, and largely unreported. That changes now.

Voice Clone Scam
WhatsApp Voice Clone Targets Market Traders
Political Deepfake
AI-Generated Video of Political Figure Goes Viral
AI Chatbot Fraud
Fake Bank Chatbot Harvests Credentials on Telegram

Seen something like this? Your report could prevent the next attack.

Report AI Harm
Our Mission

Protect people.
Not just systems.

AI safety is not an abstract research problem. In Zambia, it is a livelihood issue. When a voice clone drains a mobile money wallet, when a deepfake erodes trust in an election, when an algorithm denies someone a loan because of their postcode, that is the safety failure we exist to prevent.

Our founders have red teamed frontier models at OpenAI, including GPT-4o, Deep Research, and alpha tested DALL-E 2 through GPT-4o Mini. They have also evaluated Anthropic models from Claude 3.5 through Claude 4.6, and contributed safety work for Canva AI, among others.

Beyond evaluations, we have built our own experimental world model from scratch and are developing Red Sand, an open AI safety platform for the African continent. Our founders bring this expertise and much more back home to build the safety infrastructure Zambia needs.

Track Record

Verifiable experience.
Measurable impact.

0+
Frontier models red teamed by our founders
0%
Open research, freely accessible
0/7
Public harm reporting
Intelligence Briefing

Stay ahead of
AI threats.

Weekly intelligence briefing: top harm reports, emerging threat patterns, model safety scores, and policy updates from Zambia and beyond.

Free forever. Unsubscribe anytime.

Opportunities Database

AI safety is a team sport.

We will be opening fellowships, research positions, red teaming contracts, and community roles across AI safety. Join our database now to be first in line when opportunities go live.

01 RED TEAMING

Red Teamers & Evaluators

Adversarial testing of AI models, building attack scenarios, and contributing to safety benchmarks for frontier systems.

Coming Soon
02 RESEARCH

Researchers & Academics

Publish open safety research, contribute to alignment studies, and build the evidence base for AI governance in Africa.

Coming Soon
03 POLICY

Policy & Governance Experts

Shape AI legislation, draft regulatory frameworks, and advise governments and organizations on responsible AI deployment.

Coming Soon
04 TECHNICAL SAFETY

ML Engineers & Developers

Build safety tooling, interpretability infrastructure, and monitoring systems for AI deployed across Zambia.

Coming Soon

Apply now to join the database. Be first when opportunities open.

Join the Database
Partners & Collaborators

Trusted by frontier organizations.

OOpenAIRed Team Partner
DADeepLearning.AIEducation Partner
ZZICTAGovernment Partner
BBaseBlockchain Partner
VVercelInfrastructure
OOpenAIRed Team Partner
DADeepLearning.AIEducation Partner
ZZICTAGovernment Partner
BBaseBlockchain Partner
VVercelInfrastructure
OOpenAIRed Team Partner
DADeepLearning.AIEducation Partner
ZZICTAGovernment Partner
BBaseBlockchain Partner
VVercelInfrastructure
OOpenAIRed Team Partner
DADeepLearning.AIEducation Partner
ZZICTAGovernment Partner
BBaseBlockchain Partner
VVercelInfrastructure