The creator of a viral video purporting to show Japanese Prime Minister Fumio Kishida making explicit sexual admissions in a live news broadcast said he made it in about an hour as a “little joke.”
“I didn’t think it would create such a stir,” the man in his 20s said of the video, explaining that he used generative artificial intelligence technology to create Kishida’s voice and mouth movements.
The video shows the prime minister speaking to the camera in a live news program on Japanese broadcaster Nippon Television Network Corp. The company’s logo appears in the top right corner of the screen along with a ticker saying, “Breaking News.”
The video’s creator, whose name could not be verified, said the video did not gain much attention immediately after it was first uploaded in July this year. But when he uploaded it again in November by extending its length to around 3 minutes and 30 seconds, it went viral on social media.
The man said he conceived the idea after watching news reports on the prime minister meeting with AI experts.
Feeling Kishida showed little interest when given a demonstration in which AI software changed a meeting participant’s voice to resemble that of his own, the creator said he thought of “making fun” of Kishida by using the same tool.
He had the software learn Kishida’s voice from publically available clips before tasking it with repeating remarks that the man read out. He used AI to adjust Kishida’s mouth movements to match the voice.
It took about two hours to have the software learn Kishida’s voice, but the video-making itself was completed in about an hour, he said.
“I want to apologize to Nippon Television Network,” he said.
Generative AI can utilize vast amounts of data from the internet and other sources to generate text, images, or other media with basic prompts from a human operator.
On the back of the rapid development of the technology, many countries have grown wary of the spread of misinformation, including the circulation of fake videos of politicians on social media.
Following Russia’s invasion of Ukraine in February 2022, a deepfake of Ukrainian President Volodymyr Zelenskyy calling on his country’s military to surrender spread.
Japan’s top government spokesman Hirokazu Matsuno warned in a press conference on Monday, “Posting of false information hurts the foundation of democracy and can constitute a crime.”
Hiroyuki Fujishiro, a professor at Hosei University who specializes in social media, said, “It is difficult to see through a shrewdly made fake video at times of emergency such as during war or natural disasters.”
“As abuse of news materials leads to social confusion, there is a need for a system to clarify if content is made by generative AI and then posted on social media and other platforms,” Fujishiro said.
Source : Kyodo News