AI in the 2024 Election: Speed, Scale, and Voter Risk

概览

This Marketplace Tech episode discusses a Center for Democracy and Technology report on how generative AI appeared in the 2024 U.S. presidential election. The main conclusion is that AI did not produce the wave of deepfakes or spectacular deception many feared, but campaigns did use it to work faster and at larger scale.

The conversation with Tim Harper focuses on why the worst predictions did not materialize, including voluntary norms and fear of voter backlash. It also warns that those norms may weaken as campaigns become more accustomed to manipulated media and feel pressure to keep up with opponents.

The episode looks ahead to future elections, especially 2026, where fragmented state rules, better targeted misinformation, and AI-influenced search results could create new risks. Harper argues that public education and voter resilience need to be renewed every election cycle.

分段落总结

[00:00] Sponsor Message: Supply Chains and AI

[事实] The opening sponsor message says companies can no longer build procurement and supply chains only around cost and efficiency. [事实] GEP is described as using an AI-native platform and services to help enterprises gain visibility, manage risk, and make capital investment decisions across global value chains. [事实] The sponsor message says GEP has more than 7,000 employees across 30 offices and supports over 1,000 organizations worldwide.

[00:55] Introducing AI and the 2024 Election

[事实] Host Megan McCarty-Carino introduces the episode by saying there have not been many AI “shenanigans” in the election so far, though that could change. [事实] The episode frames the 2024 presidential race as the first major election in the new generative AI era. [事实] The host says there were major concerns that generative AI could deceive voters or interfere with democracy, but a new Center for Democracy and Technology report found that such activity had been limited.

[01:44] Campaigns Used AI for Speed and Scale

[事实] Tim Harper says campaigns mostly used AI for speed and scale, not for spectacle. [事实] Consultants told the report authors that AI’s main value was rapid, targeted messaging across multiple audiences rather than producing deepfakes. [事实] Harper says the real AI change in politics is faster persuasion, faster data analysis, email creation, outreach, and campaign strategy work.

[02:20] Why the Worst Predictions Did Not Materialize

[事实] Harper says voluntary guardrails and developing norms helped limit the use of manipulated videos in 2024. [事实] He says campaigns worried voters would penalize them if they created manipulated videos. [事实] He warns that if voluntary norms erode, campaigns may start to view manipulated content as more rewarding and less risky. [推测] The episode suggests the restraint seen in 2024 may be fragile rather than permanent.

[03:32] State Laws and a Fragmented Regulatory Environment

[事实] The host notes that more than half of U.S. states have passed laws related to AI in political advertising. [事实] Harper says many of these laws require disclaimers when manipulated content seeks to influence an election outcome. [事实] He says the rules differ across states, including the size and duration of labels, what is allowed, and the time periods covered. [事实] Harper says the lack of a federal standard creates a complicated patchwork for campaigns to navigate. [推测] Larger campaigns may benefit more from AI because they have the resources to adjust ads for different markets and legal requirements.

[04:35] AI May Advantage Larger Campaigns

[事实] Harper says larger campaigns with more resources may be better positioned to gain the benefits of AI tools. [事实] He says this continues a trend from 2024, when early adopters were larger campaigns rather than underdogs. [事实] He contrasts this with the belief that AI tools might level the playing field for smaller campaigns. [推测] The discussion implies that AI could reinforce existing campaign resource gaps instead of reducing them.

[05:29] Midterm Risks: Voting Misinformation and Search Poisoning

[事实] Harper says AI could be used to spread misinformation about the time, place, and manner of voting. [事实] He gives an example of AI-generated text messages using public voter files to make false voting information seem personally credible. [事实] He says AI could also be used to poison search algorithms so voters receive incorrect, outdated, biased, or misleading information. [推测] The main risk described is not only fake media, but personalized misinformation that uses real voter details to appear trustworthy.

[06:51] Public Resilience and Ongoing Education

[事实] Harper says there was a significant investment in 2024 in building public resilience to AI threats. [事实] Public communication included guidance about detecting AI-generated images, such as distorted fingers or poorly rendered human features. [事实] Harper warns that because 2024 did not bring the expected catastrophe, public outreach and education may receive less emphasis later. [事实] He says AI resilience should be renewed every election cycle, just as voter education about when, where, and how to vote is repeated every cycle.

[08:21] Report Link and Credits

[事实] The host says the full Center for Democracy and Technology report will be linked on MarketplaceTech.org. [事实] Jesus Alvarado produced the episode. [事实] Megan McCarty-Carino closes the episode for Marketplace Tech.

[08:39] Promo for How We Survive

[事实] A promotional segment introduces How We Survive, hosted by Amy Scott, as a podcast about climate solutions. [事实] The promo mentions geoengineering ideas including balloons sent into the stratosphere and sunshades that could dim sunlight across Earth. [事实] The promo says the show is available on podcast apps.

播客点评/总结

This episode is valuable because it moves the AI-and-elections discussion away from only deepfakes and toward more practical campaign uses: faster messaging, data analysis, outreach, and strategy. That makes the risks feel more concrete and less sensational.

A strong point is the distinction between 2024’s limited visible misuse and future vulnerabilities. The interview makes clear that the absence of a major AI-driven election crisis does not mean the issue has disappeared.

[推测] The episode is best suited for listeners interested in election integrity, political technology, AI governance, and campaign operations. Its limitation is that it relies mainly on one report and one interviewee, so it does not include direct responses from campaigns, regulators, or voters.