Grace Health Develops Ethical Framework for SRHR Information in AI-based Solutions
With Grace Health’s vision of improved health for all women everywhere and leveraging the strength of mobile technology, automation and artificial intelligence to put female health directly into the hands of women worldwide we know that we have an amazing opportunity but also a responsibility. That is why we brought together a multidisciplinary group of experts to look closer at the ethical impact of our AI, and to develop a framework for bias-reduced Sexual and Reproductive Health and Rights (SRHR) information and services inside AI-based solutions.
Why are we doing this?
At Grace Health, we are pioneering a sector that traditionally has relied upon human interaction for information dissemination. Societal norms have put women in disadvantaged and discriminatory structures. Our vision at Grace Health is improved health for all women everywhere. We believe that we can realize this vision by leveraging the strength of mobile technology and artificial intelligence (AI) to put female health directly in the hands of all women worldwide and deliver a scalable platform to the 1.9 billion women in emerging markets who own their own phone but in many cases lack access to smart and relevant health services and information. We see AI and automation as a scalable tool to overcome discrimination and obsolete social norms and to solve the current problem of accessing health services and information instantly and with discretion.
But making use of AI does not mean that we can completely disregard the impact of human interaction. Services like the Grace chatbot that use AI and Machine Learning (ML) to remedy challenges of discrimination often overlook the fact that AI itself is a tool created by humans and thus prone to biases and norms. As we at Grace Health aim to serve women at scale and to develop services for the next billion users, designing for inclusivity has always been one of our top priorities. Therefore, we decided to partner with some of the top expertise on norm critical content as well as the ethical and societal impact of AI: to dig deeper into the impact of biases and norms on our AI.
We approached the Swedish Innovation Agency, Vinnova, whose focus on norm critical innovation perfectly aligned with the work we wished to do and last fall we received a grant to carry out the work.
For the past six months, Grace Health has worked closely with the project partners to design workflows and frameworks for creating norm critical and bias-reduced educational content on the topics such as identity, sexuality, sexual pleasure, violence and more.
“With the current 250 000 women using Grace Health’s AI-powered female health assistant, Grace, and with a steadily increasing growth, we couldn’t be happier about collaborating with this knowledgeable multidisciplinary group of experts to drive development and usage of methods, tools and processes to support norm critical and norm creative innovation within AI ”
Therese Mannheimer, Founder and CEO of Grace Health
Building on the insights of this work, the larger aim of the project will be to together with our academic partners develop an ethical framework for SRHR education, information dissemination and services inside AI-based solutions.
“Women’s health and sexual education are of crucial importance to empower women across the world and to improve their knowledge and behaviour. AI-based systems can help make information available to women in a timely and personalised way, in particular in places where women would be less able to get this information due to cultural tabus, demographics or geographic distances. However, developing and using AI in a responsible way is crucial, given the importance and sensitivity of the topic. This is more than ensuring unbiased data, but also requires transparency of processes, accountability for the results and use of the system, ensuring participation and inclusion of users and stakeholders, and openness about aims and approaches used. We are happy to work towards these aims with Grace Health.”
Virginia Dignum, Professor at Umeå University, Wallenberg chair on Responsible Artificial Intelligence.
We hope our joint work can serve as an important catalyst, inspiration and framework for startups and larger companies and help them design and build norm critical AI.
“In general, society is biased, and this bias is reflected in collected data which is then used to build AI. So, bias is propagated from humans, over data, to AI systems. We have to realize that it’s kind of a double-edged sword: AI has great benefits for our society, but we also know that AI can be biased in a way that leads to prejudice and discrimination. We hope that our research can help to raise more awareness about biased AI and develop computational methods that mitigate unwanted bias so that it is not amplified”
Suna Bensch, Associate Professor at Umeå University
The project parties
RFSU, the Swedish member association of International Planned Parenthood Federation (IPPF), has a long-standing expertise in providing sexuality education and methodologies to address stereotyped norms, counteract myths and address misconceptions.
The Department of Computing Science at Umeå University, is one of the leading research departments in ethical artificial intelligence. It consists of a collective of renowned researchers with expertise on AI ethics and responsible AI, human-AI interaction, computational linguistics, chatbots and social robots. Within the project, the team is exploring the potential ethical challenges, potential threats, and how to best make sure that AI capability is well understood and applied properly.
Virginia Dignum is a professor at the Department of Computing Science at Umeå University, Sweden where she leads the research group Social and Ethical Artificial Intelligence. She is also a member of the European High Level Expert Group on AI, and Fellow of the European AI Association (EURAI). Expertise on AI, AI ethics, and human-AI interaction.
Suna Bensch is Associate Professor at Umeå University. She holds a Master’s degree in computational linguistics, and a Ph.D. in theoretical computer science. She recently completed her role as PI for the VINNOVA Vinnmer project “Integrative Human Language Technology”.
Botkyrka municipality represents great diversity with 52 % of its population having another ethnic background than Swedish. The Botkyrka municipality has a long and in-depth competence in developing inclusive and accessible educational material to people of all ages.
The results and learnings of the project will be published by the end of 2020. Stay tuned for progress updates along the way.
To learn more about the project don’t hesitate to get in touch at firstname.lastname@example.org. To learn more about Grace Health and the first-ever digital women’s health clinic designed for the next billion women online, go to www.grace.health