G20 Brasil: an opportunity to shape how we shape technology
What’s the problem? Data from others, from the world and from ourselves can be used by companies and governments in ways that affect our lives. So why don’t we have a say in how this is done?
What’s the opportunity? There is a growing movement for more democratic control of technology and space for experimentation, with cases of public interest governance of data and AI arising across the world.
Why G20 Brasil? Together, the Group of 20 (G20) members represent 85% of global GDP, over 75% of global trade, and about two-thirds of the world’s population, playing “a significant role in shaping and strengthening the architecture and global governance across all major international economic issues.” Moreover, Brasil has a long and strong history of social movements mobilisation and collective redress to push for rights.
This is blog post of a op-ed written by Maria Luciano - Research Associate at Connected by Data, IEA-USP; Vinay Narayan - Senior Manager of Policy at Aapti Institute; and, Alison Gillwald - Executive Director at Research ICT Africa. It was published by Estadao on 16 January 2024. A pdf version in Brazilian Portuguese is available here.
Data and AI conversations need to include people to serve them
As connectivity continues to be extended, artificial intelligence systems rise, and digital government infrastructures gain momentum around the world, we begin to question whether these developments are embedded with mechanisms and processes to ensure they truly serve people. How is the public interest, and the values, needs and concerns of affected communities being accounted for?
The Unified Payments Interface (UPI), India’s instant inter-bank peer-to-peer and person-to-merchant payment system is a remarkable example of what (re)thinking governance can achieve. This digital public infrastructure (DPI) managed to foster digital inclusion in the country by ending a data monopoly from the two companies that dominate payment data from the whole world, Visa and Mastercard, and while building a system that technically does not have lock-ins to a particular bank or platform.
However, since the UPI is only an underlying protocol that can be built on top of, this space has witnessed private capture with Google Pay and PhonePe (owned by Walmart) accounting for 80% of the UPI market in India in 2022. Critically, a lack of regulation on how financial transaction data can be used and shared has allowed these private companies to use transaction data to yield valuable customer insight, offer other services and, in combination with other data collected by these tech behemoths, to build market intelligence that can be used to respond and even nudge users or mould choices towards certain commercial outcomes.
In addition to this, the broadening of access to digital banking has also welcomed scams that leverage the UPI platform. By exploiting people’s lack of understanding of the digital space, fake money requests and even fake websites are used to facilitate fraudulent money transfers.
Much like in India, Pix, Brasil’s instant payment system, has also fostered digital inclusion, but has sparked an increasing number of fraud cases and harms related to the system. This has spawned a lack of trust and understanding from citizens about how the system works and the risks posed by it.
On paper, DPIs like Pix and UPI provide the ability to wrest back control of critical digital infrastructures and place them in the public domain – thereby enabling digital service delivery in a transparent and accountable manner while unlocking societal value to foster development. However, a lack of forethought regarding data use and access policies constrains the public’s ability to use transaction data for broader social good while allowing private entities to benefit from such data. Access to data is elevated in the context of redressing not only the differential impact of harms associated with data-driven technologies but also the uneven distribution of opportunities for data value creation. Artificial intelligence systems, which require vast quantities of data to be trained on, for better and improved performance, have further compounded these inequities.
There is a growing movement for more democratic control of technology. The elements of a future in which data and AI are subject to public governance are there, in many sectors and regions: residents collecting data on air quality and noise pollution to influence municipal-level decisions in Barcelona; deliberative processes to assess what the public perceive as ‘public good’ or ‘public interest’ uses of data in the United Kingdom; community-based natural resource management resulting in animal populations being restored while creating jobs and generating revenue in Namibia; access to government information on public spending enabling civil society to monitor government spending in Paraguay; a multi-stakeholder 13-member Steering Committee deciding in which cases it was legitimate to share data beyond government use in Ghana during the pandemic; data-gathering initiatives giving indigenous communities significant control over the definition and dissemination of their data in Canadá.
For the first time, the G20 presidency is held by a series of developing nations - India in 2022, Brasil in 2023 and South Africa in 2024. This creates a window of opportunity for Global South issues and integrates them into a common, public interest agenda. In this context, DPIs could prove especially important. In an environment where trends in cross-border data flows seem to be dominated by traditional powers and big tech platforms, the success of the UPI in India has seen it become a powerful tool of digital diplomacy for India, who have shared the technology with other countries including France, Singapore and the UAE.
Adopting DPI systems can provide countries with greater autonomy over data flows involving their citizens as the underlying infrastructure for crucial public services is developed, built and controlled within the country. This however must redress underlying digital exclusions and be combined with mechanisms and processes that ensure such technological interventions truly serve people without being extractive. Doing so will require active participation to not be limited to ad-hoc participatory practices involving only the usual suspects and appropriating the public debate with technical language. It will need to go beyond groups of “data subjects” and “consumers”, acknowledging the relational and public good qualities of data - non-rivalrous and can, through open data and commons mechanisms, be made non-excludable to include all affected people and communities in these conversations.
The long and rich history of collective mobilisation and redress in Brasil, India and South Africa can provide guidance on how this can be achieved. Abalobi in South Africa is one example. This initiative, designed with and for fishers, is a data and platform cooperative that enables fishers to be co-producers of knowledge and owners of the data it collects, while digitally supporting fishing communities, creating market opportunities, and better informing public decision-making. It has also helped to shed light on the labour of the women within the community.
An enabling environment is required to ensure the necessary scale and scope of these types of initiatives and the G20 presents a suitable forum to push for this. The engagement groups of the G20, such as Civil20, Science20 and Think20, allow for various stakeholder groups to contribute to the G20 process and it is incumbent on these stakeholders to make the voices of affected people and communities heard at such a forum.
Let’s not miss the opportunity to reshape our relationship with data and AI, and make everyone’s voice heard.