Chinese-American victim of deepfake porn attack by man she spurned sets up firm to fight malicious content
- Rejected man wants to humiliate her, posts fake images on porn sites, she hopes tech firm can empower victims to take on haters
A Chinese-American victim of deepfake porn at the hands of a man she had spurned has fought back by setting up a company to help victims like her.
Breeze Liu, now 28, said in April 2020, a friend told her that a nude video of her was widely circulating on an adult website.
She knew it was fake and discovered that the edited clip used her face, reported the news website people.com.
“I was devastated,” Liu was quoted as saying. “I was so humiliated and felt so alone.”
She suspected the footage was created by a man she met before her graduation from the University of California at Berkeley two years earlier.
When Liu confronted him, the man said because she had rebuffed his advances, he wanted to “humiliate her”.
At Liu’s request, the man deleted the video from the adult website. However, he later shared similar posts on hundreds of other sites.
Liu said her life was ruined by the spread of the manipulated content.
“I was trying to live a normal life, but deep inside, every second I was screaming,” she said.
“Every person who passed by me on the streets, I would wonder, have they seen the video?”
She called the police in Berkeley, California, but was told they could do nothing. They even asked if Liu had engaged in prostitution.
“Justice would not be possible for me unless I change the entire system,” she said.
Now, Liu, who had years of work experience in the cryptocurrency sector, is the founder of a small company developing free-to-use facial recognition software that will allow victims to identify fake and abusive content and remove it.
She used her own savings to set up the San Francisco-based venture Alecto AI.
The venture is in partnership with major platforms to facilitate the removal of deepfake material which came into being thanks to artificial intelligence (AI) image, audio and video generators.
Her work got her an invitation to meet US President Joe Biden last year, when he signed an executive order on AI.
“I can’t control what congress or law enforcement does. But the one thing we should all be able to control is the sovereignty of our own images online. What I’m building is a first line of defense that gives victims their power back,” said Liu.
Deepfake content has been prevalent across social media around the world for several years.
In January, fake porn images of Taylor Swift went viral, leading her legions of fans to defend her by reporting the accounts en-masse and efforts to bury the images.
In February, a Ukraine blogger complained on YouTube that her face and voice had been used by AI technology to present her as a Russian person to sell goods at big platforms in China.
A woman in eastern China’s Shandong province called the police last year after finding her pictures were used without her consent in porn videos which were released on overseas websites.
