Grace health - changing attitudes around SRHR

Important results from the project

We have democratised knowledge within SRHR through creating a norm critical AI focusing on making it more impartial. Since we realised that we should aim for bias reduction after conducting research on creating norm-critical environments, we believed it was essential to learn how and where bias occurred. Therefore we looked into Grace’s product structure and developed a framework in content creation and user interaction monitoring. We exceeded our goal, and our new improvement affected more than 600,000 women since the mid of August 2020.

Expected long term effects

We developed our framework to create a norm-critical environment that corresponded to our user’s societal development. Our in-product survey showed that 69% of respondents said they had learned new things, 61% of them agreed that Grace service is friendly, and they feel safe to ask any question without being judged. This proves that our improvements helped users gain new knowledge that can help them make informed decisions regarding sex topics. The framework also included methods detecting bias to ensure our AI and content is as non judgemental as possible.

Approach and implementation

We designed the project using an iterative process, enabling an agile and closed collaboration. It contained four work packages across four parties. Each party took the lead on their respective packages and worked closely with the core team. Running a multi stakeholder project with partners from both public, private, academia and municipality was very rewarding but also sometimes challenging due to the working culture and pace in the different parties. In the end we overcame our differences and made them our strengths which helped us deliver a successful project.

External links

Grace Health Develops Ethical Framework for SRHR Information in AI-based Solutions Our point of view on the definition of un-bias AI and it´s impossibility Grace Health Website