SDG 8: Decent Work and Economic Growth
SDG 9: Industry, Innovation and Infrastructure
SDG 10: Reduced Inequality
2. Project Details
Company or Institution
Home Lending Pal
AI and Digital Ledger Technologies
General description of the AI solution
At Home Lending Pal, we are building a unique AI ecosystem in the consumer lending space that promotes race equality, financial literacy, innovation and economic growth in a secure and tamper-proof manner. It provides a multi-sided benefit for both the lending and borrower side by providing key insights that greatly simplify and automate the largest financial purchase for individuals. For AI models, data is the “oil” that drives the outcomes and decisions made in increasingly impactful ways. If this data is not properly secured and protected from bias, the resulting outcomes from the models can result in inequality and unwarranted consequences. Further, as AI becomes more assertive in global economies, the underlying data that drives these models has increasing significance due to extensive breaches that have compromised sensitive customer information and sold to third-parties for profit. As a result, people have lost trust in data aggregators and how their personal information is being used. As these concerns continue to grow, nations are recognizing the need to enact regulation to protect user data from nefarious agents. For instance, the General Data Protection Regulation (GDPR) was enacted by EU law to enforce user data protection and mitigate risks associated with compromised data. However, the challenge posed to global organizations is how to leverage technology to protect user data and build AI models that mitigate bias and data leakage. Our novel solution focused on the lending industry provides the technological framework for substantially improving data privacy, AI bias and promoting fair outcomes in a responsible and safe construct. We leverage the latest in digital ledger and AI technologies and processes that seamlessly integrate in building a trustful data science pipeline. This results in the user controlling their personal information, which builds the delicate trust and confidence needed to deploy responsible AI at scale.
Home Lending Pal
Excellence and Scientific Quality: Please detail the improvements made by the nominee or the nominees’ team or yourself if your applying for the award, and why they have been a success.
Our responsible AI leverages innovative capabilities within the AI and machine learning space in a unique and pragmatic way that fundamentally changes how the lending industry adheres to fair and unbiased practices, that ultimately benefits society as a whole. This is achieved using a multi-faceted approach that applies the latest advancements in AI and digital ledger technology that a) greatly reduces the time required to process a loan application 2) increases data security and data privacy 3) mitigates data bias and provides human explainability. The digital ledger technology provides the borrower institutional-grade anonymity and data privacy coupled with a level of transparency in the lending process never achieved before. This is applied in a multi-sided marketplace that promotes transparency for key stakeholders including borrowers, lenders and underwriting. In regards to creating a fair and ethical AI system, we leverage multiple techniques and processes at all stages of the data science lifecycle to reduce risk and exposure to biased outcomes. This includes a) involving a diverse group of stakeholders in the planning process of the machine learning solution b) a data governance framework that detects bias from source data c) exhaustive testing of our machine learning algorithms d) iterative MLOps pipeline that incorporates model explainability for all outcomes and predictions e) guard rails when edge cases are encountered by the model that signals either retraining, data inspection, or model validation. We provide extensive documentation for all our AI and machine learning processes that can be used by auditors or compliance officers for a deeper understanding of our data and modeling approach. Furthermore, we are currently at TRL-6 and testing to achieve TRL-7 shortly with a provisional patent for our digital ledger efforts with media coverage in Forbes, local media outlets, podcasts and pitch competitions.
Scaling of impact to SDGs: Please detail how many citizens/communities and/or researchers/businesses this has had or can have a positive impact on, including particular groups where applicable and to what extent.
Our fair lending solution will have a positive impact on roughly 3 million African Americans, 2.8 million Latin Americans, and roughly 3.1 million millennials with less extensive credit history. We anticipate these numbers will continue to grow exponentially as we further penetrate the lending landscape. With respect to the overall objectives and scope of the SDG goals, our AI solution addresses the following areas: Decent work and economic prosperity. For minorities and BIPOC communities, our solution addresses a fundamental bias that has limited and hindered their ability to access high quality lending products and services. Industry, Innovation and Infrastructure. While our focus is on the lending industry our AI framework can be generalized to a host of diversified industries. Reduced Inequality. By leveraging privacy preserving technologies, data anonymization, and explainable machine learning we quantitatively reduce inequality in the financial marketplace. With respect to measurable progress towards the SDGs, we have a well-defined user feedback system and user generated content that continuously analyze and validate our value proposition. By providing explainability in our AI models, it results in a more trustful relationship that produces a collaborative human-machine teaming environment and helps facilitate the transition from traditional processes. As AI continues to progress and make increasingly impactful decisions, its mission critical to have deep insights into how the AI is making its predictions as it affects human safety and financial literacy. This framework lends itself to be deployed in a variety of applications. With respect to global impact, data privacy leveraging distributed ledger technology and secure AI provides far-reaching applicability across all UNESCO Member States. This is because we provide a sustainable solution for addressing inequalities and systemic bias that is highly scalable to grow to an international presence.
Scaling of AI solution: Please detail what proof of concept or implementations can you show now in terms of its efficacy and how the solution can be scaled to provide a global impact ad how realistic that scaling is.
In regards to impact, our lending solution has been used by a wide-variety of people seeking to purchase a new home. According to Google Analytics we have had a total of 19,000 users create accounts since we entered our public beta. 800 (8%) of those users have reported successfully becoming a homeowner and (15%) have reported financial improvement. We have secured pilots to advance our system with Flagstar Bank, The Mortgage Collaborative (234 lenders), and Happy State Bank. Further, by leveraging IBM’s cloud infrastructure, we can support rapid expansion of our services to millions of potential customers rather quickly and seamlessly. We also employ a fast agile development environment for iterative development, testing and deployment of our AI models. Another advantage of our solution is that lots of our growth has been organic and based on actual testimonies and experiences of users. We do have a strategic market plan to cater not only to the underserved communities, but all minorities and millennials looking to make arguably the largest purchase in their lives. In regards to the emergence and growth of companies using AI, we provide a blueprint for using responsible AI in a safe and ethical manner. This is a fundamental idea as AI continues to make increasingly impactful decisions in people’s lives. One way to promote this type of engagement is to open-source specific processes that can be easily replicated and applied to a diverse set of business use cases. Another avenue is collaboration with universities, research centers and other companies seeking to further the innovation is the fair lending space. Further, we continuously monitor and research latest trends and technologies in the AI space for practical application into our solution framework.
Ethical aspect: Please detail the way the solution addresses any of the main ethical aspects, including trustworthiness, bias, gender issues, etc.
From an ethical aspect, our solution provides an unbiased and fair platform for borrowers to acquire the best lending outcomes for purchasing the largest economic investment in their lives…a home. Regardless of gender and ethnic background, borrowers now have a level playing field for obtaining the best lending product available, based on their financial history and credit profile. With regards to trustworthiness and bias, we leverage multiple layers of innovative technology that keeps the borrower’s data safe and in control of how it is used. Technologies such as blockchain, synthetic data and explainable AI provide the borrower deep insights on the decision making process of lenders in a transparent way, and generate a recommendation plan to improve their financial standing. In addition, a trustful AI ecosystem is created using robust testing and simulation to ensure models perform well and avoid bias and drift. For the short-term, our focus is on disrupting the lending space using our extensible AI framework. Given this focus, we have partnered with several key constituents including The Mortgage Collaborative, Flagstar Bank and Happy State Bank to name a few. As we continue to grow our footprint in the lending domain, as a strategic objective we will seek to apply our framework in all parts of the financial transactions business from obtaining credit cards to purchasing a car. With regards to the trustworthiness of our AI solution, it is lawful and fully compliant with all applicable regulations and goes beyond data privacy and security requirements. Further, the AI solution we employ adheres to ethical principles and values. Finally, we offer a highly robust AI solution, leveraging principles from adversarial AI, rigorous testing, simulation and monitoring.