AI Safety Zambia / Est. 2026

MAKING AI
SAFE FOR
ZAMBIA.

We work to prevent catastrophic and existential risks from advanced AI by red teaming frontier models, monitoring AI harms, building governance frameworks, and growing Africa's AI safety research capacity.

Why We Exist

Advanced AI poses risks
from scams to extinction.

AI systems are becoming more capable and more widely deployed. Without adequate safety measures, they pose risks ranging from current harms (fraud, deepfakes) to catastrophic and existential threats (misaligned AGI, loss of control).

We work to prevent both by building safety infrastructure before the gap becomes a crisis.

Existential Risk from Advanced AI

As AI systems approach and potentially exceed human intelligence, ensuring they remain aligned with human values becomes exponentially harder. We work to prevent catastrophic outcomes from misaligned AGI.

Near-Term AI Harms

WhatsApp voice-clone scams, deepfake political content, and AI-driven mobile money fraud impact Zambians daily. These current harms reveal alignment failures that scale to existential risks.

Deployment in Fragile Contexts

Frontier AI systems deployed in environments with weak institutions and limited oversight reveal safety failures first. Zambia's context exposes vulnerabilities that affect AI safety globally.

Building Safety Infrastructure

We translate AI safety research into operational tools: red teaming, harm monitoring, policy frameworks, and research capacity. This infrastructure reduces both near-term and existential risks.

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

Prevent catastrophic risks.
Build safe AI for Africa.

AI Safety Zambia works to ensure that advanced AI systems, including future AGI, are developed and deployed safely in Zambia, Africa, and globally. We address both near-term harms (deepfakes, fraud, algorithmic bias) and long-term existential risks (misalignment, loss of control, catastrophic deployment failures).

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 Canva AI, among others.

Beyond evaluations, we have built our own experimental world model and are developing Red Sand, an AI safety platform for the African continent. We bring world-class AI safety expertise to African contexts, revealing failure modes that global labs miss and building the safety infrastructure Africa 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