人工智慧（Artificial Intelligence）至今也有了不少發展，從AlphaGo在圍棋界的一戰成名開始，在許多領域中AI也不負眾望取得輝煌戰果，但這位超級英雄是否有黑暗邪惡的一面呢？而當這位機械英雄壞起來成為邪惡魔王，我們人類能否有足夠的力量抵抗呢？

觀點

職場產業

AI 危害！ 細數近年邪惡AI

大陸最近一款新的「AI換臉」APP，被網友熱烈討論。只要用一張正臉照，就能輕易合成。把藝人的臉換成自己的；立刻變身大明星演戲。就連曾經任職微軟、GOOGLE的創新工場董事長 李開復，都親自體驗。不過專家提醒，這可能侵犯藝人的肖像權，臉部特徵還可能遭有心人士盜用。

【全新東森新聞APP】一手掌握全世界→https://goo.gl/Ui51Xe

#東森新聞 #EBC #最新時事

陸瘋「AI換臉」APP 一秒變臉化身大明星

Last December a Reddit user named "deepfakes" posted a pornographic video with the graphically superimposed image of Gal Gadot's animated face on a porn star's body.  In the video Gadot's face doesn't blink, but it was otherwise a reasonably convincing fake.  This precipitated a brief but highly predditublicized blitz of amateur graphic artist trolls posting face-swapped videos online. 

Now, just about anyone can be in almost anything — and there's no clear solution any serious consequences of it.

Reddit, Twitter, and other websites banned these "involuntary" pornographic videos. But "deepfakes" videos, as they have come to be known, remain online on mainstream sites more in the humorous variety. Posting videos of Nicolas Cage's face, for example, is particularly popular.   

While celebrities and public figures are rightly concerned about the implications of this technology, its roots are not in the dark corners of the internet. 

Last week in Vancouver the world of computer-generated graphics companies gathered at the SIGGRAPH conference to discuss their trade. It is an annual conference where the motion picture industry, computing giants like Google and Adobe, and leading computer science research institutions present their latest research. 

Researchers at the Technical University of Munich, along with colleagues from Stanford, presented their work to make deepfakes even more advanced. Their work is intriguing, but when an audience member asked them how they planned to confront the ethical implications of these easily replicable fake videos, nobody had a satisfying answer. It was as if the researchers hadn’t considered how the technology they were developing could be misused. 

The U.S. government thinks that’s a problem.  This year DARPA will spend more than 28 million dollars developing a way to detect and debunk deepfake videos. But even if DARPA helps funds technology that can detect a fake video, that doesn’t mean people won’t be able to make it — and distribute it quickly — anyways. 

Subscribe to VICE News here: http://bit.ly/Subscribe-to-VICE-News

Check out VICE News for more: http://vicenews.com

Follow VICE News here:
Facebook: https://www.facebook.com/vicenews
Twitter: https://twitter.com/vicenews
Tumblr: http://vicenews.tumblr.com/
Instagram: http://instagram.com/vicenews
More videos from the VICE network: https://www.fb.com/vicevideo

科技問題