The successful release of ChatGPT in December 2022 has brought artificial intelligence (AI) to the forefront of discussions in many corporate boardrooms.
The ability to generate human-like responses to almost any query with surprising confidence has fascinated many.
“AI” is an umbrella term that encompasses a wide range of technologies, including machine learning, natural language processing, computer vision, robotics, and more. Several recent breakthroughs have enabled the emergence and introduction of tools such as ChatGPT. 
The cost of implementing these models has decreased
There are innovations such as the Transformer architecture introduced in 2017, which allows AI to analyze entire sentences at once to improve understanding of context. At the same time, improvements in hardware (such as GPUs and TPUs) have made it possible to process larger datasets more quickly.
Additionally, self-supervised learning allows the model to learn on its own from unorganized text, making it easier to understand complex language patterns.
From a practical perspective, AI tools like ChatGPT are becoming more user-friendly and accessible to the general public, and non-technical users can also benefit from them. Additionally, the cost of implementing these models has decreased, making AI technology more affordable and practical.
trust the response
But can you trust the accuracy of the responses generated by AI? That largely depends on how you use it.
Large-scale language models (LLMs), such as ChatGPT, in their current form can cause “hallucinations” or produce false or misleading information when used as standalone chat tools. This occurs because LLM generates text based on patterns in large datasets without verifying the accuracy of the content.
Approximately 20% to 40% of unsecured consumer loans require manual processing of documents, creating unnecessarily high operational costs.
Their confident tone can give the false impression that misinformation is reliable. While this remains a challenge, ongoing research aims to reduce hallucinations by improving how models are trained to identify and avoid potentially inaccurate claims.
More focused applications that use AI to analyze unstructured data within a well-defined framework, rather than generating new content, make it easier to ensure accuracy and performance. As a result, tasks that were previously done manually and required hours of work, such as checking income, employment, and affordability, are reduced to seconds.
If you provide financial services, there are several important factors to consider before using AI tools. First, remember that if the tool is free, the data can become the product. AI models often learn from user interactions, so asking the right questions is important in regulated environments where personal information is handled.
LLM is particularly suited to understanding unstructured text-based data.
We ensure that your personal data is securely stored and managed in accordance with GDPR and other regulations. This includes implementing strong security measures such as encryption and access controls.
Additionally, ensure that the data is not used for training beyond its intended purpose, especially if it contains sensitive information.
Next, check whether your AI provider has appropriate certifications, such as ISO/SOC 2 standards or Ethical AI certification, to demonstrate best practices and regulatory compliance.
Finally, consider how AI will impact customer outcomes. If AI has a significant impact on decisions that affect customers, audits should be conducted to ensure fairness and detect bias in the system.
What are some specific examples of AI being used in practice in financial services today? The LLM is particularly suited to understanding unstructured text-based data. Much of the customer information in financial services is contained in customer-provided documents, leading to significant manual effort and inefficiency.
Can you trust the accuracy of the responses your AI generates? That largely depends on how you use it
Get a mortgage – All applications still require bank statements, pay stubs and tax returns. This leads to manual review queues, countless manual efforts per application, and a fundamentally substandard customer experience.
Similarly, 20% to 40% of unsecured consumer loans require manual processing of documents, creating unnecessarily high operational costs.
AI tools can streamline this process, reducing costs and processing time by up to 75% through automated verification of income, employment, affordability, and more.
Alexis Rog is the founder and CEO of Sikoia
This article was published in the November 2024 issue of Mortgage Strategy.
If you would like to subscribe to our monthly print or digital magazine, please click here.
