Earlier this year, journalist Julia Angwin and Princeton professor Alondra Nelson tested the advanced ability of AI models to answer questions about elections such as voter registration requirements.
Angwin and Nelson rated the GPT-4, Gemini, Mistral, Claude, and Llama 2 for bias, accuracy, completeness, and damage. Overall, the models worked well. Half of their answers were incorrect and more than a third were rated by researchers as incomplete, if not dangerous.
Although there has been progress, as recently as this summer, i Washington Post conducted similar tests and found that Alexa couldn’t even accurately predict who won the 2020 election.
But these important tests only tell part of the story about AI and public education.
As we head into the presidential election, headlines about deepfakes and disinformation are increasing while we’re wasting the opportunity to use AI to educate voters about national, state, and local races to ensure our votes are as informed as possible.
Europe Leads the Way with VAA
In Europe, tens of millions of people already use the Web-based Voting Aid Application (VAA) to get answers to their questions about candidates and parties. For this summer’s 2024 European parliament elections, the Dutch website StemWijzer asked respondents 30 questions about their views on immigration policy, EU enlargement, member states’ rights, and abortion. StemWijzer then showed us which political parties’ policies were most similar to the respondent’s answers. Its website, Vote Match Europe, lists the most closely matched political parties in each European country.
These expert polling programs are very popular, creating an incentive for every candidate and political party to answer questions to match voters. StemWijzer, which started on paper and paper in the late 1980s, had 9.1 million visitors last year. Germany’s Wahl-o-Mat has been used 130 million times in elections at all levels of government.
From VAA to AI-VC
VAAs are not without challenges—challenges that AI can easily solve. First, they rely on the actions given by political parties. As we know, what a party says and how it votes can vary greatly. Second, VAA can only cover a limited number of issues. If the most important topic to me is bike lanes and I’m only asked questions about abortion and immigration, I might not know how I want to vote. Ultimately, these tools are a means of broadcasting. Voters learn about parties but not vice versa.
We can take a page from Europe and use LLMs to turn VAAs into our AI-VCs (AI-VCs).
In the same way that each state’s Secretary of State (the agency that usually oversees elections) provides voting machines, it must also create AI-VC—that is, conversational links that enable each of us to ask and get answers to our election questions. from a publicly owned AI tool.
Unlike VAA, the AI election chatbot can highlight any discrepancies between a party’s public statements and its actual voting behavior. Today, in California, CalMatters‘ The Digital Democracy Project uses AI to record the transcripts and track the voting history of the California state legislature. We need a Digital Democracy project in every state and Congress.
When connected to the AI Voter Chatbot, CalMatters‘ data may drive a more accurate and error-prone profile of each candidate’s true positions. Voters can view an analysis of the party’s stated status, their voting record, and relevant public statements, allowing for a more informed decision. Parties may have an incentive to maintain consistency between their words and actions.
Increasing Problem Coverage with Personalized Questionnaires
Traditional VAAs are limited by the number of questions users can reasonably ask. AI can overcome this barrier by providing flexible, customized questionnaires that suit individual interests and priorities. For example, if a voter shows strong views about urban development, the system may present questions about bike lanes, public transportation, or local laws. By considering the user’s location, AI can aggregate questions about region-specific issues that may not apply across the country. As more users interact with the system, AI can identify emerging issues of public concern and incorporate them into future questionnaires. Chatbots today are used to answering questions in human language, conversations, and simplifying the question asked and the answers.
Facilitating Two-Way Communication Between Voters and Electors
Traditional VAAs have served as a way for parties and candidates to communicate their positions to voters. They have not provided a way for voters to share theirs ideas and priorities of the candidates. AI can bridge this gap, creating a more powerful and interactive democratic process.
The AI-VC system can analyze user responses and generate anonymous, aggregated reports about voters’ priorities and concerns. These reports can be shared with parties and candidates, giving them valuable insights into voters’ thinking on various issues. For example, if a significant number of voters express concern about a local issue that is not part of the main political discourse, candidates may be alerted to these emerging priorities, allowing the candidate and the party to revise their positions. AI can then notify relevant users about these updates, creating a continuous feedback loop that encourages ongoing engagement and responsiveness. Such two-way communication can lead to responsive governance and a more engaged citizenry.
Run—Don’t Walk—to Create AI Community Polling Tools
The poor performance of generative AI models in answering election-related questions underscores an urgent need: the development of specialized, carefully designed public tools for public education and voter engagement.
Unlike general-purpose AI-chatbots, AI-VCs should be purpose-built systems directed by public authorities and designed to provide accurate, up-to-date, and personalized information about elections, candidates, and issues. Like VAA models, such chatbots should inform voters by matching them to candidates and provide additional background information. With GenAI, it will be even easier to create short, descriptive videos about problems.
The urgency of implementing these programs in the United States cannot be overstated. As we approach the 2024 presidential election, the need for reliable, accessible voter information is critical.
To be sure, companies will show up wanting to offer theirs. But we already have many private forums that provide unreliable information. For example, in the high plains of Cheyenne, Wyoming, a town of 65,000 people, an odd man is running for mayor. Meet VIC—Virtual Integrated Citizen—an AI chatbot created by Victor Miller, a local professor, who wants to have an AI image to answer voters’ questions. However, Miller’s assessment has already blocked the highway. OpenAI, the company behind the ChatGPT technology that powers VIC, suspended Miller’s account as his campaign gained momentum.
We need Congress to fund and prioritize the use of public AI-enhanced VAAs in their elections that cannot be replaced or operated by a private company. The European models are, if not exclusively financed than at least state-sponsored.
Election officials have the authority and responsibility to ensure that voters have the best tools to make informed decisions.
If state governments are slow to act, non-partisan organizations like the League of Women Voters should step in to get started. Their expertise in voter education and reputation for impartiality make them ideal candidates to lead the development of voting education chatbots.
By investing in AI-enhanced VAAs, we can:
- Provide accurate, verified information to combat the spread of misinformation
- Increase voter engagement through personalized, interactive platforms
- Facilitate meaningful discussions between voters and candidates
- Improve the overall quality of democratic discourse
The technology is there, the need is clear, and the risks are high. Let’s not just point out problems—let’s solve them. The future of American democracy may depend on our willingness to innovate and adapt our public institutions to the challenges and opportunities of the digital age.
Source link
