by Venkat Kodumudi, Chief Technology Officer
The trust of the public in the Federal Government has remained consistently low for the past few years. To address this issue, President Biden has recently issued an Executive Order #14058, issued 13th December 2021) directing specific agencies to improve the services they provide to the public. Federal employees and citizens alike have struggled with the availability, accuracy, and interpretation of content for a range of services. Large Language Models (LLMs) have recently gained popularity as a potential solution. Unissant, as part of its Innovation Center strategy, has been experimenting with LLMs as a part of a feasibility study to see their value and usage in the Federal Government. Contrary to negative press regarding their usage, LLMs have several beneficial use cases, such as secure and efficient contextual search results and accurate translations of complex legal terms.
Recent negative press that LLM technology has garnered has led to many doubts regarding its appropriate usage. However, Unissant's feasibility study reveals the inherent benefits that controlled usage of LLMs can offer in the Federal Government setting, specifically in improving Customer / User Experience (CX/UX). Focused usage of LLMs can dramatically improve the experience of federal employees and citizens seeking services, as well as enhance their trust in the Federal Government. The controlled use of LLMs can provide accurate contextual search results, ensure secure transmission of sensitive data, and improve the quality of services provided.
One of the most common Federal Government services for LLMs to address is the search for patent information on the U.S. Patent and Trademark Office’s (USPTO) website. While the search functionality has improved in recent years, use of LLMs makes it easier by eliminating the need to know the exact context, text, and other information. Instead, LLMs can provide relevant and accurate search results, making the search process more efficient for users.
Large Language Models can also help Federal employees looking for HR and other related information by reducing the need to call or email someone. LLMs have the capability to provide timely and accurate answers, which can help employees manage their work better, etc. This provision of contextual information through LLMs saves Federal employees and citizens much-needed time and effort and contributes significantly to enhancing their experience.
LLMs can also help ordinary people who lack the know-how to understand complex legal jargon. An example of this would be someone trying to understand the terms and conditions of a Federal RFP. In this scenario, LLMs can "translate" the technical wording in layman's terms, making it easier for non-experts to understand the document's content. This feature of LLMs is a boon for common people who are eager to access Federal Government services but are intimidated or hindered by the complexity of the language that comes with the process.
Lastly, LLMs can significantly help senior citizens looking to benefit from Medicare and Social Security services to understand the application process. Senior citizens are often overwhelmed with vast information and must depend on others to navigate the complex application process. LLMs can provide clear and accurate guidance on how to apply for Medicare and Social Security benefits, making it easier and faster for senior citizens to access the services they need.
LLMs are a game-changing technology that can significantly improve customer and employee experiences within the Federal Government. True, recent events have colored some areas of public perception of the LLMs' usage. However, the appropriate and controlled application of LLMs in the Federal Government is crucial in propelling the services provided into the future. The machine-learning algorithms used by LLMs will further enable accessibility, efficiency, and accuracy in contextual search results, and provide an extra layer of confidentiality and privacy to sensitive data - all critical elements in building and enhancing trust in the Federal Government.
Unissant is committed to continuous innovation and exploration of new technologies. Taking cue from our partners in the Federal Government, we are currently experimenting with Large Language Models (LLM) at the Unissant’s Innovation Center (UIC) and are eager to explore these and other use cases with our Federal Government partners. Our aim is to develop solutions that not only meet the current needs of the Federal Government but also anticipate future requirements. We look forward to the opportunity to collaborate and innovate together, pushing the boundaries of what's possible with LLM and other emerging technologies.
