In recent months, fabricated visuals and audio clips created by artificial intelligence have flooded social media, often portraying elected officials in compromising or absurd scenarios. A digitally altered photograph showing College Park, Georgia mayor Bianca Motley Broom sharing a kiss with an unknown man spread across Facebook before being debunked, illustrating how quickly a synthetic image can reach thousands.
Similar incidents have emerged nationwide: a manipulated picture of activist Nekima Levy Armstrong was shared from a White House account, an AI-generated scene placed former New York mayor Zohran Mamdami with Jeffrey Epstein was circulated from a parody account, and Senate Democrat Raphael Warnock reported seeing deepfake videos of himself with fictitious grandchildren and a full head of hair. These examples underscore a growing concern that deepfakes—hyper-realistic synthetic media—can erode public confidence just hours before a ballot is cast.
Political deepfakes erupt across the nation
Beyond images, voice-based scams have proven equally damaging. Two years ago, Tarini Padmanabhuni’s grandfather was fooled by a deepfake telephone call that mimicked his brother’s voice, demanding a ransom. The incident spurred Padmanabhuni to launch DetectifAI, a startup that embeds detection algorithms directly into smartphones, preventing the audio from ever leaving the device.
According to an FBI report, Americans lost nearly $900 million to AI-driven scams last year, with seniors over 60 accounting for twice the losses of younger adults. Meanwhile, a June poll conducted by Rad Campaign and Women Who Tech revealed that 71 % of respondents fear deepfakes will influence upcoming elections, and two-thirds worry about AI-powered sexualization of women and children. The data demonstrates that the public’s anxiety centers more on manipulation of democratic processes than on economic disruption.
Legislative race to curb synthetic media
Lawmakers at both federal and state levels are scrambling to create safeguards. Senator John Kennedy proposed a bill obligating AI developers to embed a “kill switch” to deactivate rogue models, yet the proposal stalled after Senator Rand Paul labeled it premature. In the House, Minority Leader Hakeem Jeffries urged immediate, broad-scale regulation, calling for “guardrails to protect the American people from bad actors and bad behavior”.
State initiatives vary. Tennessee’s “Elvis Act” leverages right-of-publicity statutes to shield political figures from unauthorized voice cloning, while most states still lack explicit deepfake disclosure requirements for campaign ads. Georgia, despite several bills introduced in 2024 and 2025, has yet to enact any concrete protections, leaving officials like Mayor Motley Broom vulnerable to fabricated attacks that reinforce gendered stereotypes.
Tech firms answer with on-device detection
DetectifAI represents a new category of defense: compact AI models designed to run locally on smartphones. Unlike cloud-based services offered by competitors such as Reality Defender or Microsoft Azure AI Content Safety, DetectifAI’s SDK evaluates audio in real time without transmitting the recording, thereby preserving user privacy. The company has already licensed its technology to phone manufacturers and processes over 100,000 calls each month for financial institutions in India, where it verifies speaker identity and flags deepfake voices.
By embedding detection at the hardware level, the startup hopes to create a market advantage similar to AT&T’s exclusive role in the original iPhone launch. Early investors—including former TechCrunch editor Josh Constine—see the approach as a crucial antidote to a threat that is growing faster than legislative responses.
Meanwhile, political candidates are taking legal action. Democratic congressional hopeful Rebecca Cooke in Wisconsin issued a cease-and-desist letter to Republican Rep. Derrick Van Orden after he released AI-generated videos that falsely portrayed her in conversations with President Joe Biden. The letter cited Wisconsin law prohibiting deepfakes in political messaging, though the campaign exploited a loophole that applies only to state, not federal, candidates.
These skirmishes illustrate a broader pattern: as synthetic media becomes easier to produce—thanks to tools that require only a prompt and a consumer-grade computer—both the political arena and the technology sector are forced into a rapid arms race. Whether legislation will keep pace or whether on-device detection will become the de-facto standard remains an open question, but one certainty is clear: the battle over truth in the digital age has only just begun.
