Blog

News and developments from RD4C

New Publication

Responsible Data for Children in the Context of AI and Emerging Technology

Reflections from an Expert Conversation

Posted on 7th of March 2024 by Andrew Zahuranec, Stefaan Verhulst, Eugenia Olliaro, Friederike Schuur, Krisana Messerli

Responsible Data for Children in the Context of AI and Emerging Technology
Responsible Data for Children in the Context of AI and Emerging Technology

“Data and AI can help deliver vital services to those most in need. It can increase accessibility to education, healthcare, and other vital services for child and human flourishing.”

“Data and AI have the potential to widen existing inequalities within and across countries; there is a risk that the rich get richer and the poor get poorer.”

On February 26, 2024, the Responsible Data for Children initiative (co-led by UNICEF at Chief Data Office and The GovLab at New York University) convened a “brain trust” dinner to discuss the future of data and AI for children. Hosted by the Doris Duke Foundation and supported by UNICEF USA, we came together united by optimism that a better future is possible and a shared belief in the role that data and data technologies can play in enabling every child to reach their full potential. However, we were also united in concern that the steps that are currently being taken (or not taken) may not put us on the right path – the path to realizing the capacity of data and data technologies for social and public good to the benefit of all. 

Together, we identified pressing needs in our efforts to improve child wellbeing, along with possible solutions for the Responsible Data for Children initiative to carry forward through its work in countries across the globe.

New Tech, Old Problems

When people talk about emerging technology, there is a tendency to focus on new problems to solve with new ‘exciting’ tools. But participants seemed to agree that a better use of new technology is to solve old, protracted, difficult problems we have been struggling to solve for years; inclusion, child well-being, access to quality education, healthcare.

Many countries are more concerned with how data systems must be enhanced to improve civil registration, digitized administrative systems, and more. All these seemingly basic changes have a profound impact on the lives of children and young people. 

“It’s important to think about the basics,” said one expert. “ And the basics are finding kids at birth and following them as they grow up [...] ensuring they are protected throughout the process. Tech must build upon existing foundations to contribute to success.”

“The theme I hear from this conversation is that we should not always ‘shoot for the stars.’ We should use tech to deal with what we are struggling to do currently and should have done before instead of immediately leaping to new problems.” To stop getting distracted by ‘sugar coated shiny objects’, we should be concrete about what new technology can do to old problems.

Data as an Enabler

“What do you think are the most exciting opportunities opened up by data and technology for children who are being born today?” 

An hors d’oeuvre to brain trustees who brought to the table expertise in statistics, technology, product development, international development, and child well-being from National Statistical Offices, Permanent Missions to the UN, philanthropies, and technology companies.    

The central concept of “data is an enabler, never a solution itself” was defined early on. “Technology by itself will not do it” neither, specified one participant. While new data and technology provides exciting possibilities—for example, using low-cost audio doppler systems to identify high-risk pregnancies to improve maternal health outcomes—it is not a cure-all. 

Both data and technology need to be embedded in a supportive ecosystem—for example, trained staff to use these low-cost doppler systems, as well as structures and processes for referrals and care where needed. 

Responsible Governance

Data and technology are tools like any other, neither inherently good nor bad. There is a critical need to identify ways to use technology for social and public good and to minimize the ways it might be risky, harmful, destructive, or a driver of widening inequalities. 

Securing Transparency and Building Trust

For several participants, a major gap around data and technology systems is the scarcity of trustworthy tools stemming from the lack of understanding of what went into them. We cannot derive meaningful answers from new tools without understanding what fed them (both quality and origins of the data) or what assumptions undergird them. 

Relying on their experiences working on welfare in the United States, one participant shared that they always asked themselves, “What are the biases within the system [...] that appear objective but in fact embed norms and value judgements?” 

Efforts to build trust in new technology systems could therefore create significant harm if there were not significant changes in how they are designed, developed and deployed. This concern is particularly acute with regards to generative AI. Participants argued that “people are being given agency to use machines with blind trust. They are being granted power over something that only ‘sounds’ like it is operating correctly [because] if a machine gives us information that sounds intelligent, it must be true.” 

“We cannot let [generative AI] make decisions about children when there’s so little transparency about how results emerge.” This unease about the use of generative AI without knowing how it works and what guardrails are in place led us to a conversation around regulations and accountabilities.

Strengthening Regulations

“To put it bluntly, a lack of governance kills kids. [...] Technology is sexy, but we need to focus on unsexy stuff”, said one participant. Our laws and regulatory mechanisms have not evolved to guarantee AI tools are designed and used responsibly. To do good, we need principled and flexible rules, checks and balances that reflect our fast-evolving reality. Without them, we risk deepening inequities — as communities who have the resources to use (and produce) new technologies will use them to help themselves, while others languish.

To this point, one participant flagged the importance of self-governance and urged everyone, in particular tech companies, to help those underserved so as to reduce systemic gaps, both globally and within communities.

Similar to what Statistics Offices do, tech companies could state the origins of the data used by their technologies and include a level of confidence or margin of error in the results their systems produce. This would help companies explain how their systems generate answers or from what datasets those answers emerge. 

“The rules themselves should be shaped in such a way as to give people the most choices throughout the process. It’s like voting. You go more than once.”

Fostering Participation, Agency and Literacy

In this regard, participants emphasized the importance of engaging young people in the conversation, ensuring they have their finger in the pie. “We need to hear their voices and understand their concerns.” “Including young people in the conversation [of how tech is used] has the possibility to democratize technology.”

For their participation to be meaningful, some urged the need to make both adults and children more literate about technology. Additionally, while high-tech systems can erode our ability to critically interrogate them (a risk of turning us into passive consumers), reinforcing skills like programming and critical thinking can stymie that loss. One compared skills training to math, noting that “kids need to learn to do math themselves before they are given a calculator, otherwise they’ll only ever rely on the calculator and fail to grasp the foundational concepts.”

Meaningful agency will not be possible until these problems are addressed. 

“Tech gives tools to young people to expand their capacity. We should think about how we can give them [meaningful] agency in these conversations.”

***

As we enter our fifth year, the Responsible Data for Children initiative continues to pioneer ways to address new and emerging data challenges. The themes that emerged during discussion will guide a yearly Responsible Data for Children strategic dialogue to set priorities and select projects for the year. But doing this alone is not a piece of cake… Our next strategy session is in mid April, please do not hesitate to reach out with suggestions at [email protected]

If you’d like to stay up to date with us as we continue this journey, we encourage you to sign up for our newsletter

 

 

Photo by Brooke Lark on Unsplash 

Back to the Blog