John P. Pelissero is director of government ethics at the Markkula Center for Applied Ethics. Views are his own.
Artificial intelligence (AI) is reshaping the way we find, process, interpret, and use information in all sectors of society. AI will impact the way our representative governments operate and have the potential to enhance or weaken the bond between citizens and their elected representatives in our democracy. Achieving significant benefits from the deployment of AI in democratic governments requires its responsible use by our public officials.
From the time of our founding as the United States of America, this republic has embraced the system of representative democracy to conduct the people’s business. The founders recognized that our fledgling nation in 1787 was already too large and too populous to conduct its national government affairs through a direct democracy in which citizens would gather periodically to make decisions for the people. They were also distrustful of the possible outcomes of direct democracy by the masses. Instead, they chose a system of representatives to be selected by voters during regular elections who would meet as a Congress to discuss, debate, and decide the people’s business.
In the introduction to their 2011 book, The Future of Representative Democracy, Sonia Alonso, John Keane, and Wolfgang Merkel note that Thomas Jefferson was a proponent of representative democracy, believing it to be “…a way of governing better by openly airing differences of opinion–not only among the represented themselves, but also between representatives and those whom they are supposed to represent.” The connection between citizen and representative would be maintained through correspondence, public meetings, and dissemination of information in which representatives could know what the citizens desired and citizens would have information to judge the fairness and effectiveness of their elected representatives.
Over the past 230 years, the relationship between citizens and representatives has evolved in many ways due to the advancement of communications and technology. Personal correspondence between citizens and their public officials has seen flowery hand-written letters give way to digital messages via email and text messages. Public meetings are more remote and virtual than in person. Information is shared rapidly between citizens and representatives through mass communications and digital media. Just as the Internet transformed the way that governments operated and communicated with constituents, AI will redefine this relationship and alter democracy in rapid fashion.
AI has a promise to enhance the relationship between our elected representatives and their constituents. But it must be done in a responsible fashion.
First, consider the simple and practical changes that AI can offer to representative democracy. Citizens can use the tools of AI to research government policy, its implications, and enhance their understanding of democratic institutions. Legal, technical, and scientific information that is used in policymaking can be made more accessible to the public with the use of AI tools, such as ChatGPT, to gain understanding and improve the effectiveness of their communications with elected representatives.
Similarly, public officials can harness AI to improve the quality of communications with citizens. They can obtain direct input from constituents with more precise and timely survey tools to capture their preferences and guide the representatives’ policy positions. AI tools can be deployed to rapidly summarize constituent phone messages, emails, and social media posts on emerging issues that demand the representatives’ attention.
But AI tools should not be substituted for the constitutional roles and expectations of the elected representatives.
Second, AI deployment in the work of public officials must be done responsibly and with attention to honoring the critical relationship between citizen and representative. Voters choose their representatives through elections in which they evaluate the candidate’s strengths and character to serve as a representative “of and for” the people. In most instances, especially in the case of the U.S. House of Representatives and less in the instance of the U.S. Senate, elected officials are expected to be delegates of the people and to remain in touch with those who they were chosen to lead. Good delegates are engaged with and guided by their constituents in order to provide the best possible public service. AI cannot replace the personal connection between citizens and representatives, but we are already witnessing the perversion of this relationship with AI.
In 2023, a New York state assemblyman asked an AI tool what legislation he should be proposing to the assembly. The AI-generated result guided him to work on housing legislation in which there was an apparent gap in the law. Although it was not clear that this policy issue was either necessary or important to constituents, the representative moved forward with an AI-written bill that he introduced into the assembly.
This is an example of altering representative democracy in an unintended fashion. First, the issue to which the assemblyman gave his attention was not one that came from his constituents. Instead, he decided that this was legislation that he should work on based upon the input of one source—AI—which is not a constituent. In other words, the representative took his direction as a legislator from a technology, not from the people he was delegated to serve. Compounding this failure of representation, the assemblyman had AI write the bill that he proposed, rather than write it himself with guidance and input from staff and citizens. In both steps, the normal relationship between citizens and representatives was altered in ways that undermined representative democracy.
There are many predictions of how AI will impact society and government with an ever growing list of examples of AI’s ascendency in legislative bodies, where it can supplant the expectations of citizens in a representative democracy. The scope and breadth of the impact of AI will provide ongoing challenges to our democracy.
The only way to maintain the dignity of citizens in a digital age being written (literally) by AI, is to demand that representatives consider their constitutional duty to act as faithful delegates of the people. Using AI responsibly as an elected representative is a requirement for a healthy and sustainable representative democracy that remains attuned to its citizens.
