KEYWORDS: ChatGPT, Finance, GPT-4, BloombergGPT, tax, Chatbot.
Recently, a study by the University of Florida made a stunning revelation.
According to Analytics Insight, an Indian media authority platform, ChatGPT achieved an astounding 500% return in an investment model with its outstanding accuracy, “outperforming traditional sentiment analysis models commonly used by hedge funds”[1].
In addition, a few months ago, Bloomberg, a privately held financial,software, data and media company just launched its 50-billion-parameter Large Language Model (LLM) — BloombergGPT — built from the ground up for finance. [2]
Not to mention, financial giant JPMorgan just applied for a trademark for a product called IndexGPT, which will use AI to “analyze and select securities tailored to client needs.” [3]
Creating an “AI Warren Buffett” is not so easy in the face of unpredictable financial trends. However, exciting new applications in the financial industry have already been demonstrated by recent advances in Generative AI such as ChatGPT.
Therefore, this article will discuss how Generative AI impacts the financial industry, from the benefits and challenges of ChatGPT to its application scenarios in this field.
1. Benefits of ChatGPT in Finance
1.1 More Efficient and Productive Processes
In fact, chatbot banking is nothing new. People use them to ask questions and find the latest bank policies, but unfortunately, they don’t always get what they need.
However, ChatGPT can improve and automate financial calculations in a more efficient and human-friendly way. It can save both customers and banks a lot of time and resources.
For individual users, they can simply explain their problems in their own language and ChatGPT could come up with a solution.
For financial institutions, it is like a 24/7 financial calculator or virtual assistant for financial analysts, capable of performing complex calculations such as statistical analysis and data visualization.
People can also ask follow-up questions. The chatbot uses its large language models from NLP technology to produce human-like responses that help users better understand the topic.
In addition, financial analysis professionals can use ChatGPT to automate daily tasks such as data entry, report generation, and preparing automated responses to emails.
Trading and portfolio management are also two important areas that financial institutions deal with on a daily basis.
ChatGPT is able to use sentiment analysis of financial news headlines to provide information on stock prices and their value to provide financial analysts with a well-organized summary of stock market predictions.
It can also use historical data to find patterns in the behavior of investors and companies.
Additionally, ChatGPT can track portfolio performance and provide regular updates on the overall value of the portfolio, helping investors gain insight into which assets to buy, hold, or sell, based on market conditions and other factors.
Therefore, ChatGPT can be a valuable tool for traders to make quick trades.
Photo by Mohamed Nohassi on Unsplash
1.2 Safer Risk Management
In this section, we will discuss the following two main risks:
Risk 1: Identifying fraudulent transactions and taking immediate action to prevent losses.
Risk 2: Ensuring that your customers do not get into legal or financial trouble for non-compliance.
In 2022, financial institutions in the U.S. alone lost over $4.5 billion to fraud. [4]
For example, in the digital payment scenario, the authenticity and verifiability of the user’s payment credentials are directly related to the security and efficiency of the payment.
In the digital lending scenario, the authenticity and verifiability of the personal loan information provided by borrowers is the basis for assessing their repayment ability and risk level.
Therefore, in all types of digital financial transactions, fraud detection for credentials and documents is essential to ensure their authenticity and reliability.
Based on ChatGPT’s superior processing, analytics, and in-context understanding capabilities, ChatGPT-powered AI systems will be able to identify fraudulent transactions in real time and take immediate action to prevent losses that would be difficult with traditional methods in the coming years.
By identifying the related patterns, ChatGPT can help financial institutions focus their efforts on claims that are more likely to be fraudulent.
For example, ChatGPT can be used to analyze claims data and identify claims that contain unusual or suspicious language, which could include the use of overly technical language or unusual phrasing or syntax.
Another main risk is also a major part of the work of financial professionals who analyze various legal and regulatory documents for their clients in real-time to ensure that they do not get into any legal or financial trouble for violating laws and regulations that could result in huge losses.
ChatGPT-like financial analysis chatbots can be used to create compliance policies by providing the necessary information and a set of guidelines and generating a well-structured document that meets the required guidelines.
1.3 Better user experience
Although financial chatbots are already being used to automate answers to FAQs. ChatGPT’s capabilities can turn financial chatbots into more intelligent virtual assistants.
Imagine a customer asking a bank’s chatbot, instead of receiving a cold and robotic answer, a more human and personalized response will be easier to gain customer trust and satisfaction.
For example, India’s Tata Mutual Fund partnered with Haptik, a conversational AI platform, to create a chatbot. It reduced call volume by 70% and also increased customer satisfaction by answering basic account questions and providing financial advice [5].
With the power of Generative AI, financial institutions and individuals can now have their own personal financial assistant or advisor who can provide personalized, instant and efficient customer service 24/7.
For example, people always struggle with tax issues, if a tax chatbot powered by ChatGPT can explain your personal tax situation such as the annual income, and how much you owe to Uncle Sam, as well as how people can save their everyday taxes, it will be a game-changing AI tool.
Photo by Kelly Sikkema on Unsplash
2. Use Cases of ChatGPT in Finance
ChatGPT-like tools are essential for companies that want to stay competitive, so let’s take a look at some examples like chatbots and ChatGPT plugins.
2.1 FinChat Chatbot
FinChat is a generative AI tool that provides you with information about 750+ companies and 100+ super investors. It is basically a chatbot that can act as your stock analyst assistant. [6]
2.2 Chinchilla Chatbot
Kasisto uses deep conversational AI and financial expertise to analyze account activity and generate insights to make recommendations and resolve common service requests. It also provides a multichannel experience and proactively delivers insights you can use for future improvements. [7]
2.3 ChatGPT Plugins:
2.3.1 Public ChatGPT Plugin
What sets Public apart is the depth of its market data. It provides detailed asset prices, making it easy for investors to track the rise and fall of stocks, bonds, commodities, and more.
Public also offers a wealth of news, research and analysis. It sifts through global financial news to ensure that users stay informed about important developments that could affect their investment strategies.
2.3.2 Boolio ChatGPT Plugin
Boolio enables users to efficiently analyze global equity values using quantitative factor methods. It represents the fusion of AI and sophisticated financial analysis, providing an easy-to-use platform for understanding the complexities of global equity markets.
2.3.3 BizToc ChatGPT Plugin
BizToc is a highly intuitive and efficient tool for searching business and financial news. Whether users need insight into the latest stock market trends, economic indicators, mergers, acquisitions, or startup funding news, BizToc helps filter through the noise and deliver the most relevant and up-to-date news articles.
3. Challenges of ChatGPT in Finance
Trading market conditions change rapidly. Therefore, timely and accurate information and resources are a top priority.
As we know, the current ChatGPT can handle various things, but it is not enough for more professional business scenarios such as finance.
First of all, ChatGPT depends on the quantity and quality of the data it receives. But its outdated data and resources are based on 2021 and earlier.
Outdated and missing data also leads to another result: the difficulty for ChatGPT to understand and provide in-depth analysis of financial reports or laws.
If the data is not accurate or updated regularly, it will have a negative impact on ChatGPT.
In addition, the existence of ChatGPT’s inaccurate, ethical and biased issues can cause serious losses or even violations of the law.
maadaa.ai has covered such topics, please click the link below to read more.
Furthermore, here is an example showing how to trick ChatGPT scam chats.[8]
Image Source: https://www.dexerto.com/tech/chatgpt-study-shows-how-easy-it-is-to-detect-ai-based-fraud-2157496/
A study called “Bot or Human? Detecting ChatGPT Imposters with A Single Question [9]” tested ChatGPT-3, ChatGPT-4, Meta’s LLaMA language model, Google’s Bard, and Microsoft’s Bing. All of them failed the test.
We believe that this issue will be resolved in the future, but for now, it seems that any scams initiated by ChatGPT will be easily fooled with a simple phrase.
4. Why do you need a reliable work partner?
Thanks to generative AI such as ChatGPT, the future of finance in the coming years is promising.
ChatGPT can help financial institutions achieve more efficient and productive processes, secure risk management, optimize user experience, etc.
Therefore, ChatGPT can help financial analysts and investors to make better investment decisions.
However, its limitations are also obvious, especially its outdated, inaccurate, ethical and biased data.
Therefore, if your business is to help people to deal with financial issues smarter, faster, and more efficiently, you need a reliable AI partner to help you collect, process, and label your exclusive data, save you time, and ensure the accuracy of your training data.
maadaa.ai, founded in 2015, is a comprehensive AI data services company that provides professional data services including text, voice, image, and video to customers in the AI industry.
From AI data collection, processing and labeling, to high-quality AI datasets and dataset management, maadaa.ai helps customers to efficiently collect, process, and manage data, and conduct model training for rapid and cost-effective adoption of AI technologies. [9]
Image Source: maadaa.ai
5. Commercial datasets from maadaa.ai
5.1 Chinese Bill Dataset (MD-OCR-002)
A diverse bill in different scenarios. Data collecting equipment are phones, cameras, and tablets.
Chinese Bill Dataset (MD-OCR-002) covers over 10 types of commercial receipts and invoices used in mainland China, including flight tickets, train tickets, hotel receipts, general tickets, taxi receipts, quota invoices, value-added tax invoices, toll receipts, coach tickets, and others.
Over 20 labeling categories include types, provinces, quality, codes, invoice date, company/certificate numbers, tax numbers, phone numbers, car licenses, IDs, boarding time, drop-off time, price, mileage, wait time, surcharge, service charge, and receipts, etc.
Reference List:
[2] https://www.bloomberg.com/company/press/bloomberggpt-50-billion-parameter-llm-tuned-finance/
[4] https://www.unite.ai/chatgpt-is-coming-to-finance-so-lets-talk-about-the-risks-and-rewards/
[5] https://www.unite.ai/chatgpt-is-coming-to-finance-so-lets-talk-about-the-risks-and-rewards/
[6] https://wgmimedia.com/what-is-finchat-chatgpt-for-finance/
[8] https://www.dexerto.com/tech/chatgpt-study-shows-how-easy-it-is-to-detect-ai-based-fraud-2157496/
