BetterNow’s mission is to make the world more generous. And we intend to do that while keeping a high standard for data ethics.
For BetterNow data ethics is about responsible and sustainable use of data.
We want to do the right thing for people and society. Our solutions and use of data must be designed, so they first and foremost are benefitting humans.
We consider data-ethical concerns to be more far-reaching than mere compliance with and adherence to legislation, which is the bare minimum.
Our principles that guide data ethics in our software development are:
Individuals are always at the centre of what we do and should be the primary beneficiary of data processing. We will encourage our customers and partners to act with the same "individual first" principle.
We will incorporate European privacy principles in the development of our applications. We will ask for consent and make sure it is informed. Our data architecture will always be built with privacy safeguards in mind.
We give individuals the possibility to determine the scope and length of the data processing happening. We build to ensure data portability and are happy to help individuals take control over their data.
Our data processing activities and automated decisions must make sense for the individual. We present and communicate why we process data in the way we do.
We will design systems that provide opportunities for feedback, give relevant explanations, and possibilities for appeal. When we employ machine learning algorithms, we ensure that they are under appropriate human direction and control.
When we process data, we pay special attention to how it affects vulnerable people. This is especially true when we employ machine learning.
Vulnerable people are particularly vulnerable to profiling that may adversely affect their self-determination and control or expose them to discrimination or stigmatisation, for example, due to their financial, social or health-related conditions.
We commit and adhere to the principles set out in The Toronto Declaration on “Protecting the right to equality and non-discrimination in machine learning systems”
BetterNow uses machine learning for generating personal recommendations for individual fundraisers. We take great care in ensuring that these machine learning processes do not cause harm.
We therefore:
The above is how we take our responsibility in making human rights due diligence on our machine learning algorithms, as set out in The Toronto Declaration.
Following the above process, we have changed how our machine learning algorithm works, in order to not stigmatise less financially well-off people, by removing the recommendation to increase their self-donation.
Read more about how we came to this conclusion by reading this article on self-donations in P2P fundraising.
You can find an easy to read guide on how our machine learning algorithm works, and a post on how we developed the model together with data scientists.