Sharing data between organizations can help address some of society’s biggest challenges. There is no question that working with data in any fashion requires careful consideration of privacy and security risks, but there are clear benefits to be gained when done appropriately.
For example, a group of research institutions and medical clinics recently came together to create a central data repository of information about patients suffering from amyotrophic lateral sclerosis (ALS), a progressive neurodegenerative disease that affects nerve cells in the brain and the spinal cord, as part of a project called Answer ALS. That dataset can now be used by the same organizations to improve their artificial intelligence research to advance treatments for ALS – more rapidly and with greater efficacy than they could have achieved on their own.
And yet, much of the potential that can be realized through data sharing in similar scenarios remains untapped – in part because the tools for collaboration are often immature or non-existent. Organizations that wish to share data often have to confront the reality of having to spend months or even years negotiating and drawing up contracts to govern data sharing arrangements, for example, which can be a time consuming, expensive and laborious process.
Many people and organizations are doing great work to tackle this problem. There are many data use agreements and resources readily available through an Internet search. We’ve created a set of links to many of them here.
At the same time, we think there is an opportunity to do more. We want to help make it easier for individuals and organizations that want to share data to do so. Often, agreements for broad data sharing scenarios are unnecessarily long and complex. We also think there is an important role for agreements that limit rights to computational use for AI. Further, we think the state of the art on data sharing for proprietary and private data sets is changing rapidly and the terms available publicly could be improved and better explained.
We’ve developed three data use agreements for community review and input. We have drawn from precursors, but are also advancing some new thinking drawn from our own experiences in working on data sharing projects. Hopefully you will see some of this experience come through, particularly in the annotations.
Going forward, our aim is to work with interested stakeholders to improve these agreements and to offer additional ones that cover a wide range of data sharing scenarios. Please take a look at the initial drafts and let us know what you think!
Data innovation is a complex topic, but we should work together to make the terms used more streamlined and defined. We hope these drafts are a step in this direction.
#opendata #AI #responsibleAI