Data is firmly at the center of our information society. With our online interactions, sensors, social networks and machines generating a multitude of digital traces of everything we do, companies and, increasingly, governments have turned to automated machines to make sense of this abundance of data through algorithms, artificial intelligence and other elements of the Fourth Industrial Revolution.
As highlighted in our recent publication on the bright and dark sides of data-driven decision-making for social good, there are both profound benefits and significant challenges, as massive new streams of digital data from users are analyzed, mined and used to understand and improve the society we live in.
These challenges are not just technical: there are critical social, ethical and legal questions that must be addressed. New developments in algorithms and AI affect not only the world of data science, but also our economic, legal and social structures. As the possibilities of big data, algorithmic decision-making and other innovations continue to expand, so too do the challenges raised by algorithmic discrimination, biased datasets, fake news and other new harms.
Addressing these issues requires not only data scientists, but lawyers, ethicists, advocates, policymakers, members of civil society and other stakeholders. How can these professionals, citizens and other stakeholders become more aware of and willing to engage with the data and algorithms that are increasingly part of their daily lives, in order to help prepare our institutions and frameworks for this new landscape?
Data and Algorithmic Literacy
A core priority of our work at Data-Pop Alliance is the focus on advancing global “data literacy,” which we define in a white paper as “the desire and ability to constructively engage in society through and about data.” Improving data literacy requires deepening specific capacity and specialized skills, building bridges and blending knowledge from different academic disciplines (i.e., statistics, social science, data science, ethics, etc.) and fostering a culture around data, algorithms and new technologies that incentivizes empowerment and inclusion.
In the last two years, we have designed and implemented a number of trainings and partnerships to spur practical data literacy skills among stakeholders in various contexts, notably:
- In partnership with the UN System Staff College and with support from the Hewlett Foundation, we have been rolling out our Professional Training Program on Big Data and Development, with a pilot workshop at MIT Media Lab, and two official workshops with UN, government, NGO and civil society representatives in Colombia and Kenya.
- In partnership with the Vodafone Institute for Society and Communications, we launched the digitising Europe initiative, a series of European stakeholder debates and expert dialogues on big data and its impact and implications for society, which took place in Belgium, Germany, Ireland and Spain.
- A series of training workshops for World Bank and ECLAC staff on specific topic areas at the intersection of big data and development, as well as for participants of the World Data Forum and Build Peace conference.
Described in greater depth in our workshop toolkit, these activities focus on creating diverse, multidisciplinary spaces in which participants can develop practical skills for translating societal problems into questions that data can address, identifying applicable data science tools to solve specific problems, assessing ethical and privacy implications across various big data applications, and forming cross-sector partnerships. Participants learn best practices from recent examples of big data projects and join teams with diverse skill sets to co-create next-step projects and prototypes.
Increasingly, we have been working on a similar framework for algorithmic literacy and how people can learn more about the algorithms that surround them, as a part of leading the capacity building and community engagement components of the Open Algorithms (OPAL) project pilots in Colombia and Senegal.
With Great Data Comes Great Responsibility: But Whose?
Ultimately, fostering new literacies with citizens and professionals outside of the data science world must lead to and be part of a larger inclusive discussion with companies, civil society and governments on addressing the range of challenges related to big data-driven innovations. Designers of algorithms and big data analytics must be accountable for the potential harms associated with these new innovations. In a recent report on algorithmic accountability in developing countries that we wrote in collaboration with the Web Foundation, we underline that achieving algorithmic justice in responding to fake news and other complex forms of algorithmic decision-making requires a range of technical, ethical, policy and knowledge gaps to be addressed. To implement solutions requires recognition that there is a shared responsibility of all stakeholders: algorithmic system designers, legal and regulatory authorities, public interest groups and users.
The theme of Data and Responsibility for the 2017 Data for Good Exchange (D4GX) comes at a critical time, as several communities of researchers, experts and companies around the world are joining together to reflect and evaluate the social, ethical and legal challenges presented by artificial intelligence, including Artificial Intelligence Now (AINow) (Meredith Whitaker and Kate Crawford), Data + Society Research Institute (danah boyd), the Algorithmic Justice League (Joy Buolamwini), and the Ethics and Governance of Artificial Intelligence Fund.
We’re excited to continue as part of the D4GX community and join others on the Program Committee who are involved in these discussions. D4GX continues to be an essential, multi-disciplinary convening at the intersection of data and public good. Look out for panels, presentations and ongoing discussions between social scientists, policy wonks and data scientists on data literacy, algorithmic accountability and more to be shared at Bloomberg’s 4th annual Data for Good Exchange on Sunday, September 24, 2017.
This post was written by David Sangokoya, Research Manager at Data-Pop Alliance, a global coalition on big data and development created by the Harvard Humanitarian Initiative, MIT Media Lab and Overseas Development Institute. Special thanks to Gabriel Pestre and Natalie Shoup for early comments.