Benjamin Shultz

Disinformation & Artificial Intelligence Researcher


Curriculum vitae




Enhancing Local Government Administration With Large Language Models: A Test Using ChatGPT


June 08, 2023

In the realm of public administration, the emergence of large language models (LLMs) has brought about transformative possibilities. These advanced AI models, such as ChatGPT developed by OpenAI, possess the capacity to process vast amounts of information and generate human-like responses. With their potential to revolutionize various aspects of public administration, these LLMs offer numerous benefits for officials and citizens alike.
Integrating LLMs into public administration facilitates streamlined public communication. By leveraging virtual assistants powered by these models, governments can provide round-the-clock support, addressing citizens’ common queries and guiding them through various processes. This improves citizen satisfaction while alleviating the burden on government staff, allowing them to focus on more complex tasks.
LLMs optimize service delivery by offering accurate and timely information on services, regulations, permits and public resources. Citizens can access personalized guidance through chatbots and virtual assistants, simplifying their interactions with the government. Automating routine tasks frees up resources, enabling government staff to concentrate on strategic initiatives that drive community development.
The data analysis capabilities of LLMs empower public officials to make informed decisions. By leveraging these models to analyze real-time data, governments can gain valuable insights, perform predictive analysis and develop evidence-based policies. This data-driven decision making enhances the effectiveness and transparency of public administration, resulting in improved service provision and resource allocation.
LLMs contribute to promoting accessibility and inclusivity in public services. Through multilingual support, accommodating various accessibility needs and user-friendly interfaces, these models ensure that citizens from diverse backgrounds can access government services without barriers. Moreover, they enable the participation of marginalized communities in decision-making processes, fostering equity and social justice.
The machine learning capabilities of LLMs allow for continuous improvement and learning. Governments can refine and optimize their services by monitoring user feedback, analyzing common inquiries and updating the models accordingly. This adaptability ensures that the models stay up-to-date with evolving citizen needs, resulting in a high level of user satisfaction.
The integration of LLMs in public administration represents a transformative leap in public service delivery. By harnessing the power of these advanced AI models, governments can streamline communication, enhance service delivery, make data-driven decisions, promote accessibility and continually improve their operations. As LLMs become integral to public administration, they pave the way for a more efficient, citizen-centric approach to governance, ultimately benefiting both officials and the communities they serve.
Well, that sounds convincing, right? A large language model—ChatGPT itself—wrote the above essay!
This column is rather unorthodox, however, it accurately displays ChatGPT’s incredible abilities, and why it has transformed the world in the months following its November 2022 release. Moreover, in identifying the many potential advantages of its implementation in public service, ChatGPT begins to unpack several issues facing state and local governments today.
ChatGPT is able to understand and respond to queries ranging from the mundane to literal rocket science. As governments continue to struggle maintaining adequate levels of staffing in state capitols and city halls across the country, it is becoming more difficult for them to address the needs of their citizens, on issues as simple as fixing potholes, or as complex as processing zoning exemptions. ChatGPT can help with these issues and more—and in trials conducted around the globe, it already has. From Japan, to the halls of Congress, ChatGPT has increased lawmakers’ efficiency in ways like summarizing long, cumbersome legislative texts, drafting constituent response letters and taking over monotonous administrative tasks.
Additionally, given ChatGPT’s ability to process data, it is entirely conceivable for a government agency or body to input a dataset, prompt ChatGPT to conduct some degree of quantitative analysis and leverage its computational abilities to address issues relating to budgets, staffing and capital resources. And with OpenAI’s latest ChatGPT update—GPT4—being able to produce images, graphs and charts, these possibilities appear even greater than before.
However, before governments writ large begin widely using ChatGPT and other LLMs in their day-to-day work, a number of pressing issues will need to be addressed by OpenAI and other tech companies. Between the data scraped from the internet to train LLMs, to the fact that the they are continuously trained on the prompts and replies it is fed—and any proprietary or confidential data therein—there is legitimate concern that these practices violate portions of the GDPR, CCPA and associated online consumer privacy laws. Moreover, ChatGPT and other models have been documented to produce false or inaccurate information made to sound convincing and legitimate, adding to an online ecosystem already rife with mis- and disinformation.

These issues will require collaboration between tech companies and governments at all levels to address. Until they are ironed out, public agencies at all levels must tread carefully in implementing LLMs to enhance their work. However, these challenges are not unique to LLMs—from telephones, to computers, to 5G, they have been faced before. As these novel technologies continue to evolve, and legislative frameworks governing their use catch up, LLMs have tremendous potential to address many of the critical issues impacting governments today, ideally making public services more responsive and efficient in the years to come.




Follow this website


You need to create an Owlstown account to follow this website.


Sign up

Already an Owlstown member?

Log in