by Tom Yun
Click here for updates on this story
TORONTO, Ontario (CTV Network) — The world needs to be “alert” as artificial intelligence technology improves, allowing for more realistic-looking deepfakes, a UN adviser says.
Deepfakes are media (usually video or audio) that are manipulated by AI to falsely depict what a person said or did that never happened in real life.
“A digital twin is essentially a replica of something in the real world…a deepfake is a mirror image of a digital twin, meaning that someone has created a digital replica without that person’s permission, usually with malicious intent. For some purpose, usually to deceive someone,” California-based AI expert Neil Sahota, who served as an AI adviser to the United Nations, told CTVNews.ca by phone on Friday.
Deepfakes have been used to create a variety of fake news content, including one showing Ukrainian President Volodymyr Zelensky telling his country to surrender to Russia. Scammers also use deepfakes to endorse fake celebrities. In one instance, an Ontario woman lost $750,000 after watching a deepfake video of Elon Musk that appeared to promote an investment scam.
In addition to fraud and fake news, deepfakes are also widely used to create non-consensual pornography, Sahota said. Quebec last month sentenced a man to prison for creating synthetically generated child sex abuse images using real pictures of children on social media.
“We listen to celebrities, and it really can happen to anyone, and deepfakes really started with revenge porn,” he said. “You really have to be vigilant.”
Sahota said people should watch closely because misappearing video or audio could be a sign of manipulated media.
“You have to be vigilant. If it’s video, you have to look for weird things like body language and weird shadowing. If it’s audio, you have to ask…” Are they saying things they usually say?” Do they seem out of character? Is there anything wrong? ’ he explained.
At the same time, Sahota said policymakers need to do more when it comes to educating the public about the dangers of deepfakes and how to spot them. He also suggests we need a content verification system that uses digital tokens to authenticate media and sniff out deepfakes.
“Even celebrities are trying to find ways to create trustworthy stamps, some kind of tokens, or authentication systems, so that if they are engaging in some kind of non-face-to-face engagement, they have a way to verify it. will be,” he said. . “That’s what’s starting to happen at the UN level. How do we authenticate conversations and videos, for example?”
PLEASE NOTE: This content is subject to strict local market embargoes. If you share the same market as the contributors of this article, you cannot use it on any platform.
ctvnews.caProducers@bellmedia.ca
