5

Final Thoughts and Next Steps

The workshop concluded with an open discussion focused on what should come next. Barbara Mittleman, who led this discussion, said she heard from all four breakout groups that there is an appetite to continue the discussions on ethics, data, and international collaboration. What was unclear to her, however, was what the end game would be, what this group would want to accomplish by continuing the discussions that started at this workshop and the previous one, and who else should be involved in further discussions.

With regard to who else should be included, the responses from the workshop participants included social scientists, psychologists, experts in privacy and cybersecurity, educators who are involved in teaching and developing data science curricula, field-based researchers who face the issues discussed at this workshop and who come from disciplines other than data science, experts from a range of disciplines from other countries, representatives of public advocacy groups, program officers from the National Science Foundation, National Institutes of Health, and other federal agencies who are funding research, and legislators and policy implementers, or at least someone who is familiar with the legislative process. This last suggestion was debated, with some wondering if it is too early to bring legislators and policy implementers into the discussion and others saying that it is important to involve those from the policy world at the beginning of the discussion, both to inform and sensitize them to the issues and to learn from them what is possible and what is not in the policy world.

A participant noted that he was at another meeting on artificial intelligence several months earlier and there were two members of Congress who were already thinking about artificial intelligence and how to regulate it. His hope is that this community can inform such efforts and ensure that any legislation considers the individual and their libertarian rights. As a counterpoint, another participant said it is also important to balance the rights of the individual against potential harms to society or the environment. Finding the proper balance will have to be an important part of the conversation going forward.

A workshop participant who has analyzed many international research collaborations at the national policy level noted that many Asian countries also regard bilateral research collaborations as significant collaborative forms with their partner countries, and each will be different depending on particular national and

research interests in a given country. Another participant stressed the importance of making data relevant for developing countries if the goal is to promote the sustainable development goals. He also spoke about the new type of data that will be needed to meet sustainable development goals and answer questions such as how much water will be needed to produce the additional food required to feed the world’s population and how much the new demand for minerals will affect water availability, for example. Such transdisciplinary data will require a different framework.

As to what the end game should be, one workshop participant who is involved in the policy space said she would like to see this community develop a set of best practices for data stewardship and ethics in the next 2 years and to move the discussion into the public domain. Another participant said he would like to see the discussion lead to new relationships being built that would help institutionalize new norms around ethics and data across disciplines and countries in international research collaborations.

It could be argued, said one participant, that data are shifting power in the United States away from the government toward corporations, who not only have the capability of getting the data but also to store them and convert them into knowledge. In other countries, such as China, the concern is that the reverse is true and that the government will accrue more power over individuals and corporations. Therefore, he said, it is important to keep in mind the political context of the spaces in which this field works going forward. He also stressed the importance of making sure that the data science field, along with the rest of science, technology, engineering, and mathematics, works harder to bring in people from populations that are currently underrepresented. In that context, a participant noted that the discussion needs to be more than just about data science, which is only one piece of the research enterprise.

Several participants spoke about the need to include ethics and human rights in the technology design process, rather than deal with issues related to ethics and human rights as an afterthought. Another participant made the point that data scientists tend to be doers, not deliberators, which makes it important to get these challenges out in the domains now so they can help shape these discussions. As this participant put it, if this effort is all talk, by the time the community figures it out the coders and data scientists will be very far down the road.

To conclude the workshop, one participant noted that National Science Foundation and National Institutes of Health program solicitations include a requirement for a data plan. In that regard, it might be useful to require an ethics plan as a straightforward way of having researchers think about data and ethics. Another participant added that if an ethics plan is required, some group is going to have to define what goes into an ethics. It was suggested that the American Association for the Advancement of Science Coalition for Science and Human Rights could be involved in this effort. It was then noted that there needs to be some mechanism to make sure that researchers follow through on their data and ethics plans, as well as funding and mechanisms to provide the databases and servers needed to follow through on data management plans.

The meeting concluded with promises of future communications regarding next steps and thanks to the staff, participants, and organizing committee.

This page intentionally left blank.