# ClarityAILab: safe, private AI for children and organisations

> ClarityAILab builds AI that is safe for children, private for organisations and efficient for the planet, and publishes the governance behind it.

URL: https://www.clarityailab.com/

- [Discovery, creativity and confidence, safely](https://www.clarityailab.com/clarity-kids/): Studio Clarté is a multilingual AI companion for ages 6 to 12: questions and answers, stories and practice, behind a safety filter, with no ads or tracking.
- [Remove the busywork from your supply chain](https://www.clarityailab.com/clarity-enterprise/): MaiMai structures operational information and runs sourcing and purchasing workflows, with approvals and audit trails, on the tools you already use.

## What is ClarityAILab?

ClarityAILab is a research-led AI company founded by PhD researchers. It builds two products: **Clarity Kids**, a multilingual AI learning companion for children aged 6 to 12, and **Clarity Enterprise (MaiMai)**, which uses AI to remove the busywork from supply chain sourcing and purchasing. Every product follows the same three commitments: safe for children, private for organisations and efficient for the planet.

## How is ClarityAILab different?

- **Custodianship over ownership.** We treat data as entrusted to us, not owned by us. Enterprise information and family moments stay under your control.
- **Safety by design.** Safeguards are built in, so our systems protect children and respect adults by default, without complex settings.
- **Efficiency as a responsibility.** We favour smaller, targeted models and route work to low-carbon cloud regions, so each interaction carries a lighter footprint.
- **Governance in the open.** Our policies are published in full and reviewed by an independent governance reviewer.

## Why Efficiency is a Moral Choice

AI has a physical cost. We see radical efficiency as part of our ethical obligation to people and planet.

Data centres use water for cooling, electricity for processing, and generate CO₂ emissions. We design our systems to do more with less—using smaller, targeted models and a proprietary approach to reducing token usage, so each interaction carries a lighter footprint.

We avoid unnecessary model size, data movement, and retraining. Smarter design, less waste, the same high-quality experience. We also use innovative design and cutting-edge technology to reduce environmental impact without losing quality.

**Research & Evidence**

- [Energy and Policy Considerations for Deep Learning in NLP (Strubell et al., 2019)](https://arxiv.org/abs/1906.02243)
- [Making AI Less "Thirsty": The Environmental Footprint of AI (Li et al., 2023)](https://www.cell.com/joule/fulltext/S2542-4351(23)00365-3)
- [MIT Tech Review: Training a single AI model can emit as much carbon as five cars](https://www.technologyreview.com/2019/06/06/239031/training-a-single-ai-model-can-emit-as-much-carbon-as-five-cars-in-their-lifetimes/)

We use these tools with our own families and businesses. [Download Corporate Charter (PDF)](/docs/governance/ai-governance-charter.pdf)