- Around the world, 1.7 billion people still lack access to a formal bank account.
- Digital innovations, such as AI, are helping to increase financial inclusion.
- With the appropriate infrastructure, data-sharing environment and ethical framework, AI can democratize financial services.
The volume of data generated globally is expected to increase by a whopping 530% – from 33 zettabytes in 2018 to 175 zettabytes by 2025. AI has the power to transform this data profusion into financial inclusion. We recommend various guidelines that the public and private sector can take to harness the potential of AI while mitigating the risks posed by this relatively new technology.
Digital identification systems for financial inclusion
Governments and regulators have collaborated in building new digital identification infrastructure that reduces the cost of reaching the last mile user. India has been leading the charge on this front by creating foundational digital pipelines for its “India Stack”. Prime examples include Aadhar, a 12-digit unique ID issued by the Indian government, and UPI (United Payments Interface), an interoperable, mobile-first payment system regulated by the Indian central bank.
While Aadhar has enabled bank account access by giving 1.3 billion people a trusted ID, UPI has made cross-platform, instant payments possible for the very first time. Together, these applications have spurred financial innovation by massively reducing transaction costs and enabling user data to be captured and shared on a massive scale.
Have you read?
The Indian fintech ecosystem – one of the most thriving in the world – would not be the same without these core digital platforms that enable an open and free market.
Data portability and its application to the lending and financing sector
Once infrastructure is laid, data is the fuel that enables financial service providers to extract maximum value from it. Without this fuel, fintech will be unable to travel far and without data portability, there will be minimal fuel. Data portability refers to the user's rights to easily transfer personal data from one organization to another. In practice, this requires that data no longer be stored in any one organization’s silos. Rather, it should be owned by users who can share it with other service providers. The underlying principle of this data sharing framework is that user-generated data is a public good that can, and should be, a source of competition and not competitive advantage for a single party.
This data-sharing principle is exhibited in India through the account aggregator (AA) framework. AAs are designed to be “consent managers” who enable users to share their data – on a consent basis – with other regulated entities such as banks, insurers, etc. This approach is beneficial to the lending and financing sector, particularly for the credit starved SME sector.
How is the Forum helping governments to responsibly adopt AI technology?
The World Economic Forum’s Centre for the Fourth Industrial Revolution, in partnership with the UK government, has developed guidelines for more ethical and efficient government procurement of artificial intelligence (AI) technology. Governments across Europe, Latin America and the Middle East are piloting these guidelines to improve their AI procurement processes.
Our guidelines not only serve as a handy reference tool for governments looking to adopt AI technology, but also set baseline standards for effective, responsible public procurement and deployment of AI – standards that can be eventually adopted by industries.
We invite organizations that are interested in the future of AI and machine learning to get involved in this initiative. Read more about our impact.
Approximately, 90% of the country’s SMEs lack access to formal credit. With the AA framework, these small businesses can ensure that necessary documentation is made available to lenders who in turn can better underwrite and serve small businesses. Through a thoughtful system design, the AA framework enables consent-based sharing of user data, lowers transaction costs, and enables financial inclusion.
Guidelines to develop ethical AI
As important as infrastructure and data are the AI capabilities themselves determine fintech’s penetration. While a plethora of data opens up new opportunities, the beneficiaries are limited to people for whom data is available. If the underserved are not represented adequately or the underlying algorithm is biased, AI can end up replicating social biases. This possibility is heightened by the fact that currently, there are no laws to regulate AI algorithms. In such an environment, we urge private sector players to take matters into their own hands and not wait for regulation to ensure that their application of AI is equitable. To this end, we offer three tips:
1. To reduce the risk of bias, increase the diversity of your teams
Researchers at Columbia University have found that while the biggest reason of biased predictions is biased data, the demographics of engineers also play a role. While individually everyone was approximately equally biased, homogeneous teams compounded individual biases. Thus, the more diverse the team, the better equipped it is to spot and reduce biases.
2. Ensure that your AI algorithms are explainable to others, including experts and laymen
Researchers at the Oxford Institute have developed the notion of “counterfactuals”: rather than a full-blown explanation, counterfactuals offer the minimum set of conditions that would have led to an alternative decision. Knowing these conditions can create a pathway to inclusion by empowering users with the information they need to change their current state.
3. Engage experts to audit your algorithms
Various frameworks exist to help those who want to incorporate ethics in their development cycle. Rather than reinventing the wheel, bring onboard creators of, and experts on, these tools. If expert engagement is not feasible, ensure that human intervention is not completely removed from your decision-making. While AI algorithms are getting better, human judgement remains more sophisticated and should be used to audit and design more equitable algorithms.
With the appropriate infrastructure, data-sharing environment, and ethical framework, AI can be harnessed to bring financial services to more people – particularly the poorest. This possibility must be turned into a reality.
For more information visit: corporate.payu.com