Catch tech harms in your projects.
An AI-powered tool to check for potential human rights impacts in your tech, D4D and digital programming.
HOW IT WORKS
Plain language in. Report out.
Tech Harms Check runs as an MCP server. It connects to any LLM (e.g. Claude, ChatGPT, Co-pilot). You describe or upload your tech project, it asks what it needs to know, and the tool does the rest.
STEP 01
Upload your project
Paste or upload your project info. No formal structure, no checklist, no prior human rights knowledge required.
STEP 02
Answer a few questions
The tool extracts the aspects that matter, compares them against our knowledge base, and asks only what it needs to know.
STEP 03
Receive your report
A concise PDF that flags what the potential tech harms and human rights impacts could be and what mitigation strategies might be needed. AI generated content is clearly flagged.
THE KNOWLEDGE BASE
A knowledge base, curated by practitioners.
Tech Harms Check is underpinned by a knowledge base and structured framework: tech + data × sector × context × people → potential human rights impacts → mitigation strategies.
The knowledge base has been developed by digital rights experts and practioners. Your LLM uses the Tech Harms Check knowledge base to map, score and explain the tech harms in your project. Key elements of the knowledge base and methodology will be open.
EXCERPT SCORING MATRIX
WHO IT'S FOR
Built for the people, who make the call.
Program & project managers
Planning or implementing a tech or D4D project or intervention. Need to understand what could go wrong.
Donors
Reviewing proposals that include technology components. Need a quick read on where the tech harm exposure lies.
Practitioners
Working in human rights, tech, and development. Want to strengthen the underlying knowledge base with field evidence.
SOLUTION
Plugging the gap.
Tech Harms Check has been developed to make AI-generated human rights diligence checks more accurate and reliable. In the process, it aims to solve common issues we've heard from the field.
CLARITY
Limitations and disclaimers.
Tech Harms Check is a first step check. It tells you where potential tech harms and human rights impacts might lie and what you could do to minimise these.
Tech Harms Check is not:
— A full scale human rights impact assessment
— A legal assessment or seal of approval
— A shortcut past expert judgement.
Tech Harms Check involves AI and AI gets things wrong. Always double check and interrogate responses especially when you are using them to inform your decisions.
We are designing Tech Harms Check to be a responsible AI tool. Tech Harms Check incorporates privacy by design, with all data being anonymous and encrypted. The tool is transparent about its content. It sign-posts which content comes from the knowledge base and which content is AI-generated. Key parts of the knowledge base will be made open. With its expert-developed knowledge base, the tool aims to counter AI hallucination.
CUSTOMISE
Fit to your context.
We've designed Tech Harms Check (and the knowledge base) to be adaptable. It has a human rights focus but it can be customised to use your own resources and include other dimensions or lenses – such as human-centric, governance, peoples with disabilities, or gender. Contact us to explore what’s possible.
ACCESS
Launching Summer 2026.
Tech Harms Check is a prototype is still being developed. It is being showcased at various events during Summer 2026. Access is by invitation. Reach out if you want to try it, contribute or find out more.
CONTACT
Get in touch.
Sign up to receive updates on the Tech Harms Check and its launch.
CREATED BY
SUPPORTED BY