Introduction
For most of their history, financial firms have been operating under large regulatory bodies that hamper change and make technological innovation difficult. More recently, the emergence of financial technology (FinTech) companies has put a spin on this narrative. Innovators are seeing and seizing opportunities in the finance sector to apply machine learning to improve lending practices, automate customer service interactions, and ensure transaction security, resulting in a better and cheaper customer experience. FinTech companies applying these newer technologies are creating more efficient companies and are beginning to siphon customers and market share from legacy banks. Some of these companies are beginning to operate as decentralized finance (DeFi) companies, which allow any customer with internet access to operate outside of standard regulations. This includes nearly instant and cheap money transfers, accumulating interest directly payable to the customer, and complete transparency of assets. In response to these innovations, legacy banks have invested billions of dollars into developing expansive intellectual property (IP) portfolios they hope will allow them to compete for these customers through technical innovation.
Machine Learning in Banking and Finance
Machine learning brings revolutionary changes in the banking industry, from daily customer interaction to how data is collected to fuel these innovations. As competition between legacy banks and FinTech companies heats up, the following topics are at the focus of machine learning innovation.
Lending and Consumer Products
Perhaps the most significant utilization of machine learning in FinTech has been applications for improving credit risk assessment. Improvements in credit risk assessment allow institutions and individuals to borrow more appropriate amounts of money at rates that reflect the borrower’s ability to pay it back. Square recently received patent US 11144990 titled Credit Offers Based on Non-Transactional Data, which outlines a machine learning method to analyze merchant transaction data and data associated with the merchant’s payment provider to forecast future revenue and extend credit based on the process.
Payment Security
As payment networks worldwide are growing globally, the need to authenticate and process payments grow exponentially. In response to this, legacy banks and FinTech startups utilize machine learning and filing patents to improve the process speed and security. For example, in 2020, Capital One received patent US10706422 titled Voice Recognition to Authenticate a Mobile Payment, which uses an image from the point of sale and natural language processing (NLP) to match a user’s voice from the point of sale to existing user voice data to determine if a transaction is authentic. In addition to this, PayPal received a patent for US10977654, Machine Learning Engine for Fraud Detection During Cross-Location Online Transaction Processing. This patent outlines a system and method that use machine learning techniques such as k-means clustering to process transaction data and determine if it is safe.
Fraud Detection and Anti Money Laundering
Trust is at the center of deciding where to keep hard-earned money for many. Banks know this, and as their networks expand their online presence, the occurrences of fraud and online scams have increased dramatically. The FTC reported that in 2020 there were more than 2 million fraud reports with damages totaling over 3.3 billion dollars. In response to this, FinTech companies like Hong Kong-based Austreme have developed patented technology to detect and respond to fraud more quickly and accurately. In June 2021, Austreme received a patent in Hong Kong for HK30036226 Card Testing Fraud Detection System, a device designed to receive transaction data for internet payments such as time of payment, IP address, and card number and determine if it is safe. Machine Learning in FinTech has also been used to create technologies to combat money laundering and terrorist financing. Fintel Technologies is a FinTech startup that recently received a short-term patent for US0176226 System, Method and Computer-readable Medium for Utilizing a Shared Computer System, which allows financial institutions to securely share information under provisions of the US Patriot Act Section 314(b). This shared data is then tailored to a customer’s needs and then fed into machine learning algorithms that customize a model for the customer. FinTech is not only revolutionizing machine learning applications, but the processes used to collect data, train, and test data as well.
Blockchain
As blockchain technology continues to evolve to create speed, security, and energy efficiency improvements, patents in this space have been dominated by legacy banks. Bank of America currently owns over 50 blockchain patents; however, they offer no blockchain services. JP Morgan currently owns blockchain patents, most notably US US 10,755,327, Distributed Ledger Platform for Vehicle Records. This patent enables JP Morgan to create a distributed ledger that tracks VINs for cars that they finance for dealerships. The bank already provides floorplan lending, allowing dealerships to borrow against their inventory. JP Morgan blockchain leads Christine Moy hopes that this patented technology will be applied to car financing and another financing, including heavy industrial equipment.
Trading
IP can be the difference between winning and losing billions of dollars per minute; however, many trading firms and market makers choose to keep their IP secret and not patented. A large deal of the IP created in high-frequency trading (HFT) is primarily algorithmic and not patentable in the United States, and attempting to patent parts of lucrative trading processes can reveal information that is otherwise nearly impossible to replicate. Citadel Securities, Two Sigma, and Jane Street Capital are 3 of the largest market-making firms and have combined to receive 0 patents since 2015. While non-patented machine learning undoubtedly plays a role in market making, there are patents for machine learning-based processes used in trading. US0311815 Stock Market Prediction Using Natural Language Processing outlines a process that uses NLP to intake and process information from newspapers and the internet to predict prices in the stock market. This utilization of machine learning has been used since 2001 to try to create profitable trading strategies.
FinTech IP Landscape
Legacy Banks
While traditional banks focus on protecting their ideas through patents that will pave their path to the future, newer FinTech companies are taking a vastly different approach. In general, legacy banks such as Bank of America, Softbank and Capital One are leading industry efforts to develop diverse IP portfolios. For these banks, many of these patents are not for instant use, but to lay the groundwork to protect the innovations they plan to roll out over the next decade. However, former Bank of America Senior VP Michael Wuehler, whose name appears on 8 of their blockchain patents, has said that the bank’s patents are meaningless, and the only purpose they serve is to create media coverage that the bank is innovating.
DeFi, Blockchain, and Fostering Innovation
Contrary to legacy banks, newer DeFi, Blockchain, and FinTech companies believe patents and IP litigation can hamper product development and innovation. In 2020, 25 Fintech and DeFi companies, including Square, MicroStrategy, and Coinbase, came together to form Crypto Open Patent Alliance (COPA). COPA is a non-profit community of companies that agree to remove patents as a barrier to growth and innovation in the space. COPA requires its members to pool their crypto and blockchain patents and only use them defensively as part of efforts to grow the cryptocurrency ecosystem. Membership is open to individuals and institutions of all sizes, and membership fees apply after one year and depend only on institutional revenue. Basing membership fees on revenue creates an environment that provides equal protection for both well-capitalized companies such as Meta companies and young startups such as Kraken. In addition to COPA, other institutions such as the LOT Network and OpenInventionNetwork (OIN) aim to leverage open-source technology and decrease patent litigation, which they conclude can hamper innovation.
Conclusion
Banking and finance play a critical role in our everyday lives, and machine learning innovation leads to vast improvements for companies and customers alike. Startups and legacy companies are not only utilizing advanced technologies and innovation in machine learning to improve their bottom lines but also to improve customer experiences and product offerings. While it is easy to believe that operating in a primarily regulated industry can mean limited technological innovation, it’s clear that as competition levels up in any industry, innovation will follow.
Editorial Team at Lumenci
Through Lumenci blogs and reports, we share important highlights from the latest technological advancements and provide an in-depth understanding of their Intellectual Property (IP). Our goal is to showcase the significance of IP in the ever-evolving world of technology.
