ChatGPT Rejected 250,000 Image Generation Requests of US Election Candidates
ChatGPT rejected over 250,000 requests to generate deepfake images of U.S. presidential candidates in the run-up to Election Day.
In a blog post, OpenAI claimed it had blocked hundreds of thousands of requests to generate DALL-E images of candidates in the month before the 2024 U.S. presidential election.
According to OpenAI, it rejected an estimated 250,000 image-generation requests involving President-elect Donald Trump, Vice President Kamala Harris, President Joe Biden, Minnesota Gov. Tim Walz, and Vice President-elect JD Vance.
OpenAI says that the refusals were due to safety measures that the company put in place before election day.
“We’ve applied safety measures to ChatGPT to refuse requests to generate images of real people, including politicians,” OpenAI writes in the blog post published on Friday.
“These guardrails are especially important in an elections context and are a key part of our broader efforts to prevent our tools being used for deceptive or harmful purposes.
“In the month leading up to Election Day, we estimate that ChatGPT rejected over 250,000 requests to generate DALL-E images of President-elect Trump, Vice President Harris, Vice President-elect Vance, President Biden, and Governor Walz.”
In the run-up to election day on November 5, there were concerns that AI-generated images and deepfakes could affect the presidential campaigns.
In September, California passed a new law that makes it illegal to create deepfakes related to the 2024 election — the toughest law on political AI-generated content in the U.S. yet.
California Governor Gavin Newsom signed a bill which makes it illegal to create and publish deepfakes related to elections 120 days before Election Day and 60 days thereafter. It also allows courts to stop distribution of the materials and impose civil penalties.
Governor Newsom’s aggressive laws come after he condemned Elon Musk, the owner of X (the platform formerly known as Twitter) for sharing a misleading AI-generated video of Vice President Kamala Harris in July.
However, a lawsuit was soon filed against the state of California by a content creator who creates AI-generated, parody videos — including the one of Harris that was shared by Musk.
The complainant, who goes by Mr Reagan online, says the new California laws censor free speech and allow anybody to take legal action over content they dislike.
Image credits: Header photo licensed via Depositphotos.