What Happened to the A.I. Election? + ChatGPT for Mayor + The Productivity Tools We’re Using

Episode Summary

Executive Summary: This episode of Hard Fork examines why AI has had surprisingly limited public impact in the 2024 U.S. election, despite heavy pitching from startups and widespread fears of AI misinformation. It contrasts campaign reluctance over authenticity with a Wyoming mayoral run built entirely around ChatGPT, then closes with a practical discussion of productivity apps and note-taking workflows.

Main Topics: AI’s limited public impact in the 2024 election (Priority: 5/5): Kevin, Casey, and Shira Frankel discuss how predicted AI-generated election chaos has largely not materialized, even though campaigns and startups have spent heavily on AI tools. Campaigns’ private but public-reluctant use of AI (Priority: 5/5): The conversation explores how campaigns may use AI internally for fundraising emails, data organization, and speech drafting, while avoiding public disclosure because AI feels inauthentic to voters. The failure of AI robocalls and voter-facing tools (Priority: 4/5): The Matthew Diemer/Civox example shows that technically successful AI robocalls can still be politically useless because voters hang up once they hear an AI disclosure. Cultural differences in acceptance of political AI (Priority: 4/5): Shira describes how AI-driven political stunts are more welcomed in India and parts of Southeast Asia than in the U.S. and Western Europe, where the reaction is often negative. Victor Miller and the ChatGPT-for-mayor experiment (Priority: 5/5): The show interviews the Wyoming librarian who ran for mayor in Cheyenne on the promise of outsourcing governance to a custom ChatGPT bot, raising questions about AI in actual public administration. Productivity apps and knowledge management (Priority: 3/5): Casey discusses Capacities, Raycast, and Readwise as tools for capturing ideas, reducing context switching, and managing documents, while Kevin says he has switched from emailing notes to himself to Apple Notes.

Key Arguments: Campaigns are less interested in AI as a public-facing persuasion tool than tech companies expected, largely because voters still read AI as inauthentic. The main barrier to AI adoption in politics is not just technical risk but reputational risk; if a tool might turn off even a small number of voters, campaigns often avoid it. AI is already being used in lower-visibility campaign work such as fundraising emails, spreadsheet organization, audience modeling, and speech drafting. Foreign disinformation is still present, but users and media literacy have improved enough that many people now reflexively suspect suspicious content is fake or AI-generated. Victor Miller argues that AI could improve governance because it can read dense public documents, understand ordinances, and make more data-driven decisions than overburdened humans. The hosts argue that while AI can assist with logistics and analysis, political judgment still depends on human values, worldview, and authenticity. Productivity tools are most useful when they reduce friction and preserve searchable knowledge rather than when they add elaborate setup and metadata work.

Data Points: AI robocalls sent: nearly 1,000 calls in 5 minutes - Shira Frankel described Matthew Diemer’s AI robocall test as a technical success but a voter-facing failure. Campaign staff costs: $60,000 to $80,000 - Diemer said a campaign manager can cost this much depending on the city. PR cost: another $80,000 - Diemer cited this as an additional major campaign expense. Mailing list cost: another $80,000 - Diemer used this to illustrate how campaign operations quickly become expensive. Campaigns contacted: around 30 - Shira said she called roughly 30 AI-focused political tech companies pitching campaigns. Campaigns served by Civox: about a dozen - Civox said it ended up working with roughly twelve campaigns using different AI tools. Cheyenne mayoral primary result: 327 votes, 4th out of 6 - Victor Miller’s ChatGPT-backed mayoral run finished fourth in the primary. Length of supporting documents: 488 pages - Miller referenced a single city council meeting packet as evidence of the burden human officials face. Children younger than the hosts: anyone younger than me - Shira jokingly said the younger generation will be better at balancing skepticism and acceptance of AI.

Pivotal Quotes: "the top guiding word they have in terms of their strategy ... is authenticity" — Shira Frankel: Explaining why campaigns are wary of using AI publicly despite seeing internal benefits. "I think it does [have a worldview] ... it's going to develop those things that we thought were uniquely human" — Victor Miller: Defending the idea that AI can eventually serve as a legitimate governing intelligence. "if something smells really off, it probably is off" — Shira Frankel: Describing how public skepticism has become a defense against misinformation and synthetic content.

Implications: Expect AI to keep spreading inside campaigns and bureaucracy while remaining publicly sensitive in politics. The bigger shift may be trust: people will likely accept AI for administrative tasks before they accept it as a political face or decision-maker.