Can AI Technologies Help Expand Credit Access

Despite the proliferation of digital technologies in daily market transactions, about 45% of Latin America's population do not use bank accounts, and credit access for individuals and firms is limited.

Technologies based on Artificial Intelligence (AI) can help change that. AI can efficiently incorporate big data into credit risk assessments, allowing financial institutions to profitably expand their customer base. While AI's reliance on manipulation of detailed personal data may raise consumer protection risks, it can also help financial institutions become more inclusive.

Financial Data Revolution

AI's ability to use alternative data in credit risk assessments is potentially revolutionary. Traditional credit scoring models rely heavily on formal credit histories, which many Latin Americans lack-not least, the estimated 250 million people not using traditional banking services. AI-powered models can make a difference. They have the ability to analyze diverse data sources, such as mobile phone usage, social media activity, and utility payments, to gauge creditworthiness. Studies from developing economies in Asia have shown that including data on mobile and social media usage in credit assessments improved the accuracy of loan default predictions by up to 20-30%, opening up credit to those without formal credit histories.

In Latin America, AI-driven credit scoring has been instrumental in reaching micro, small, and medium enterprises (MSMEs), as well as underserved households. About 23% of MSMEs in the region have difficulty accessing traditional bank loans due to a lack of collateral or credit records. AI-based credit scoring used by banks and fintechs have helped address these gaps by utilizing alternative data such as transaction histories and behavioral data. Brazil's Creditas, for instance, has used AI models to assess over $1 billion worth of loan applications from people without traditional credit scores, approving a significant share of loans that would typically be denied by conventional banks. Fintechs like Cashea extended short-term Buy Now Pay Later (BNPL) loans to millions of consumers in Venezuela.

The entry of big tech companies into Latin America's financial sector has further driven the adoption of AI in credit services. Online retail platforms Mercado Libre and Rappi are allocating credit using AI-powered systems that evaluate millions of data points to create sophisticated credit models. Mercado Libre, for instance, has used its e-commerce platform to offer over $1.4 billion in loans to small businesses and individuals, using purchasing behavior and payment histories to assess credit risk. Research from the Bank of International Settlements (BIS) has shown that almost a third of borrowers served by Mercado Libre in Argentina would have been unable to access credit from a traditional bank. Rappi's AI models incorporate data from its network of over 200,000 couriers delivering online orders to customers across the region. In the process, it has made lending decisions and expanded credit offerings at a rate that traditional banks struggle to match.

Complex Risk Management

AI's role in expanding credit access is not without challenges, however. Significant regulatory concerns exist around data privacy, algorithmic transparency, and potential discrimination. Brazil's AI strategy, responding to concerns that models could inadvertently exclude women, lower-income people and racial minorities, includes a focus on ethical AI use and data protection. Chile has also incorporated data privacy standards into its digital strategy to support responsible AI adoption. With the region's digital payments market projected to grow by 20% annually and AI playing a significant role in managing the accompanying credit risk, such efforts are essential.

Integrating AI into credit assessments poses technical issues. Developers are building advanced systems for data standardization and verification. The performance of AI models depends heavily on data consistency and the inclusion of diverse demographic information. Studies have found that AI models that factor in non-financial data can improve credit access by 40% for individuals without a credit history. But biases can emerge when data are skewed. That is often the case in regions where men are more likely than women to own smartphones or have household bills associated with their names.

AI applications to finance also raise privacy and security concerns. The vast amounts of personal and financial data used by AI systems make breaches and misuse particularly costly. With Latin America experiencing a rise in cybercrime-often targeting financial services-the need to safeguard sensitive data has never been more urgent. AI models often rely on complex, opaque algorithms, making it difficult for regulators and users to fully understand how decisions are made, and potentially contributing to a lack of transparency and accountability. Furthermore, as AI systems become more integrated into financial services, access to account information may become vulnerable to hacking. This is especially worrisome in countries like Brazil and Mexico, which have seen significant increases in digital payments but lag in cybersecurity infrastructure. Ensuring robust data protection laws, as well as adopting AI ethics frameworks, will be critical in mitigating these risks and ensuring that AI's role in expanding credit access does not come at the cost of security and trust.

Developing Enabling Factors

The future of AI in expanding credit access in Latin America looks promising, with ongoing efforts to enhance digital infrastructure and AI adoption gaining traction. Venture capital investment in the region's tech startups surged to $15.3 billion in 2021, compared to $4.1 billion in 2020, largely driven by the increased adoption of AI-driven financial solutions. Governments and private companies are investing in AI talent development, with countries including Colombia and Brazil leading initiatives to train more data scientists and develop regulatory frameworks that promote AI use in financial services. Today over 50% of fintech investments in the region are directed toward financial inclusion. AI, with its ability to leverage alternative data and innovative technologies, has the potential to expand credit access to underserved populations and reshape the financial landscape, providing millions of people and businesses with the credit needed to unlock their potential. While challenges remain, Latin America's proactive approach to AI adoption, coupled with regulatory efforts to address ethical concerns, offers a credible pathway toward a more inclusive financial system. As AI continues to shape the future of finance, the region stands poised to bridge the credit gap and foster economic opportunity for millions.