Original Source Here
How Scared Should We All be About Deepfakes Fooling Us
Barack Obama called Trump a Dipshit, or did he?
Last week I was flicking through TikTok and came across a Tom Cruise video talking about cleaning. I scrolled quickly past, as I can’t stand the man; sorry if you are a fan. Then I discovered that this was a deepfake, a subject I was utterly ignorant on. It appears that Cruise isn’t the first person to be the victim. Nancy Pelosi, Mark Zuckerberg, and Barack Obama have all been victims of these.
Deepfakes, although disturbing, are also amazingly clever; talented individuals use their skills to make these videos to fool many. If you have ever used face-ageing software, you have used similar technology; deepfakes are the next evolution.
How do you make a deepfake?
Deepfakes are a form of AI (artificial intelligence) developed in 2017. The technology uses a branch of AI called deep learning. These neural networks are, in essence, machine learning and work similarly to how a human brain works. It is a form of technology that is used in films. Samuel L Jackson was aged in the Marvel series by twenty-five years using this technology.
Artificial intelligence effectively learns what a source face looks like at different angles to transpose the face onto a target, usually an actor, as if it were a mask. Tremendous advances came through pitting two AI algorithms against each other, one creating the fakes and the other grading its efforts, teaching the synthesis engine to make better forgeries.
Software packages such as FakeApp and DeepFaceLab have made this technology available to all, with minimal technological skill. You don’t even need complicated equipment. Any standard computer with a good graphics card would work.
How to spot a deepfake
Some of these videos are created by amateurs, and these look fake; others, where professionals have been involved, are highly realistic. Some of the earlier ones use videos with low resolution. This is not the same for all of them though, take a look at the Tom Cruise one that fooled me.
There are, however, signs to look out for even on these versions. Look for visible changes around a persons face, known as artefacts. Is there any blurring or flickering in the video? One of the most significant clues is the eyes of the subject. In 2018, US researchers discovered that deepfake faces don’t blink normally. No surprise there: most images show people with their eyes open, so the algorithms never really learn about blinking. The eyes can also seem to work independently of each other.
Concerns with deepfakes
Not all this technology has been used for a laugh or to advance a movie; some have more sinister applications. Many celebrities have had their face swapped into porn films. The AI firm Deeptrace found 15,000 deepfake videos online in September 2019, nearly doubling over nine months. A staggering 96% were pornographic, and 99% of those mapped faces where from female celebrities to porn stars.
Deepfake technology is being weaponised against women. — Danielle Citron, Professor of law at Boston University.
As this technology advances could we be looking at fake political statements. Videos could discredit a possible candidate through things they have never said or places they have never visited. Could we all be fooled watching fakes give speeches?
A non-existent Bloomberg journalist, ‘Maisy Kinsley’, has a profile on LinkedIn and Twitter, is thought to be a deepfake. Another LinkedIn fake, ‘Katie Jones’, claims to work at the Center for Strategic and International Studies but is believed to be a deepfake created for a foreign spying operation. Last March, the chief of a UK subsidiary of a German energy firm, paid nearly £200,000 into a Hungarian bank account after being phoned by a fraudster who mimicked the German CEO’s voice. The company’s insurers believe the voice was a deepfake, but the evidence is unclear.
As technology advances, these are concerns we will all face. Most political campaigns are won as much on social media as they are in person. The same government that could be faked will then be responsible for legislating against these videos, not that this will stop everyone. Governments might be dabbling in the technology too, as part of their online strategies to discredit and disrupt extremist groups or make contact with targeted individuals.
Facebook last week banned the uploading of deepfake videos that could mislead people. Facebook is one social media network; it is clear that TikTok is a haven for this type of content. For now, check the reliability of anything you see before sharing it. Help reduce the spread of fakes and always be suspicious of anything you see or hear on the internet.
It is fair to assume that deepfakes are here to stay and could cause some significant mischief. When you remember that Tesla stock prices crashed when Elon Musk was seen smoking a joint on a live web show, think what a deepfake Elon Musk could do. It is safe to say that plausible deepfakes could manipulate stock prices and voters. Let us hope they don’t get so convincing that they convince a president, with nuclear weapons, that his counterpart in another country declared war.
Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot