Tuesday, 31 October 2023

THE GOOD, THE BAD, & THE DEEPFAKE: Q & A GUIDE


 

The Rise of Deepfake Pornography is Devastating for WomenCNN

The recent news banner above exposes the growing problem of deepfake pornography which uses artificial intelligence to create fake sexual videos of women without their consent. It reminds me of my ATABAY article The Great Unfairness I wrote in the thick of the fiercely divisive 2022 presidential election campaign, excerpts of which I am reposting below:

“There is a video of you having sex on the internet. You do not remember being with this person because it never happened. Others are watching the video online, too.

“The video is unfamiliar because it is a deepfake – an “ultrarealistic fake video” where your face is superimposed onto another person’s body through the use of artificial intelligence software. And someone has published it online for the world to see. This is not science fiction.

“All one needs is a computer, a decent graphics card, the FakeApp program, hundreds of pictures of the desired person, and a few hours. A tutorial is easily accessible online.

“Deepfakes are not only going to be used for self-gratification, but also have the potential to be used to extort, humiliate, harass, and blackmail victims.” (Deepfakes: False Pornography Is Here And The Law Cannot Protect You by Douglas Harris)

The noted victim then was Aika, the daughter of former Vice-President Leni Robredo.

Senator Leila de Lima said Aika’s deepfake was “so 2016.” That year, coming at the height of the hearing of her illegal drug-related charges, de Lima also fell victim to an accusation of having a sex video with her security aide. She stressed then:

“Demonizing women has been a go-to, and desperate political strategy in the past 6 years. Galawang sindikato talaga. Matitigil lang yan kapag natanggalan na ng sungay ang mga nagtatangkang bumalik sa kapangyarihan at yung mga nagtatangkang manatili.” (It’s what the syndicates do. It will only stop once we remove the horns of those who want to return to and maintain their power.)

Created in 2017, a deepfake then took hundreds of images of a person’s face to create a convincing video -- today, it will take just one or two images. User-friendly deepfake apps are now available for iPhones. Taking dirt cheap commission, deepfake creators can create explicit videos of a customer’s favorite celebrity, ex, or political targets like Aika or de Lima.


To thrash out this complex subject, we turn to an expert we had invited to Tech Talk.

ME: Welcome everyone to Tech Talk. Today, I’m glad to have with us a special guest, Mr. AI, a tech expert, who will share his knowledge and insights on deepfake technology. Thank you for being here, sir.

MR AI: Thank you for having me on your show. I hope I can enlighten you and your viewers on this provocative and controversial subject: deepfake technology.

ME: Absolutely, we’re excited to dive into this topic. Let’s start with the nuts and bolts. What are deepfakes?

MR AI: A pretty good start. Deepfakes are synthetic media creations made through artificial intelligence and deep learning algorithms.

ME: I’ve heard about both the bad and the good sides of deepfake technology. Can you elaborate on both?

MR. AI: Sure. On the creative side, deepfake technology has breathed new life into the following industries, such as:

Entertainment and art. It can create realistic and immersive content for movies, video games, fashion, and e-commerce. For example, the Beatles' surviving members Paul McCartney and Ringo Starr performed with the late John Lennon, bringing him back to life through his voice in their new song "Now and Then".

Education and research. It can create educational and informative content. For example, a project by Samsung AI Center and Skolkove Institute of Science and Technology used deepfake technology to animate famous artworks, such as Mona Lisa and Portrait of an Unknown Woman.

Social activism. It can raise awareness and inspire action for various causes and issues. For example, a campaign by Malaria No More used deepfake technology to make David Beckham speak nine different languages, urging people to join the fight against malaria.

ME: That’s uplifting, but what about the other side?

MR AI: Fair enough. Unfortunately, there’s a darker side to deepfake technology which can be used maliciously such as:

Pornography and harassment: Deepfakes can be used to create non-consensual pornographic images or videos of real people, usually celebrities or private individuals. They can cause psychological harm, reputational damage, and online harassment for the victims. For example, a study showed that 96% of deepfake videos online were nonconsensual pornography.

Fraud and identity theft. They can be used to impersonate someone’s voice, face, or identity and trick others into giving money, information, or access to valuable assets or data. For example, a CEO of a UK energy firm was tricked into transferring $243,000 to a fraudster who used deepfake technology mimicking the CEO's voice, bypassing the biometric security system, such as facial or voice recognition.

Misinformation and propaganda. Deepfakes can be used to spread false or misleading information, such as fake news, hoaxes, or rumors. They can influence public opinion, incite social conflicts, or manipulate elections. For example, in the ongoing Israel-Hamas war, using deepfake technology, some videos have been manipulated to show events that did not happen or to support one side of the conflict.



ME: Yikes. How are deepfake videos created?

MR. AI: Here are some steps:

Choose a tool. Online tools, such as Deepfake Web or AI studios, have features, limitations, and pricing.

Upload your videos. You need two videos: a source and a target. The source is one you change; the target is a reference. For example, you make a deepfake of yourself as Tom Cruise: you need your video as the source; Tom Cruise video, the target.

Train the AI. The tool uses the AI algorithm to learn the facial features and expressions of both videos and create a model that can swap them which can take from a few minutes to many hours depending on the video quality and length, and tool speed.

Generate the video. Once AI finishes the training, it generates a new video that combines the source and target videos. You can preview, edit, download, or share the generated video.

ME: Fascinating and instructive. But it’s unsettling to think about the negative impact of deepfakes especially on pornography. How are researchers and organizations combatting this technology?

MR AI: Fortunately, there are ongoing efforts to detect and combat deepfakes. Researchers are developing AI-based solutions to identify anomalies in media, and metadata analysis can help spot discrepancies between real and fake content. Also, media literacy and awareness campaigns are empowering individuals to recognize and respond to deepfake threats.

ME: Privacy is a huge concern, and we must address it. Could you tell us more about the privacy and ethical concerns surrounding deepfake technology?

MR AI: A pivotal question. Privacy is indeed a significant concern.  Deepfakes can be created using anyone’s likeness without their consent, which raises ethical questions. Whether it’s for entertainment or malicious purposes, respecting an individual’s privacy and consent should be a top priority.



ME: Legal and regulatory measures are crucial in addressing this issue. What’s the current legal landscape when it comes to deepfakes?

MR. AI: Some countries have introduced laws and regulations to mitigate potential harm from deepfake technology. However, legislating and enforcing laws in the ever-evolving digital realm remains a challenge. The legal framework must evolve alongside technology.

ME: As we wrap up our discussion, can you give us a glimpse of the future of deepfake technology?

MR AI: Looking ahead, we can expect advancement in detection and prevention methods. However, just as the technology evolves, so do the capabilities of those who misuse it. Staying informed and vigilant is the key to navigating this ever-evolving digital landscape.

ME: And finally, any parting words for our viewers?

MR AI: Sure. In this age of deepfake technology, our perception of reality is under constant scrutiny. We must understand the technology, recognize its potential for good and bad, and actively engage in discussion about its implications. Technology is a tool, and how we use it, defines its impact. Remember, stay curious, stay vigilant, and stay informed.

Thank you for having me.

ME: Knowledge is power, and as Uncle Ben famously told Peter in Spider-Man movie: "With great power comes great responsibility."

Thank you, Mr. AI, for joining us today on Tech Talk and shedding light on the complex world of deepfake technology. It’s been an eye-opening conversation. Thank you, viewers, for watching Tech Talk. Stay tuned for more interesting and relevant topics in our next episodes. Have a great day!


Content put together in collaboration with Microsoft Bing AI-powered co-pilot

Head photo courtesy of Adobe Stock

Video clips courtesy of YouTube

No comments:

Post a Comment

WATCHING MS. SARA: PREDICTIONS, CONTROVERSIES, AND FALLOUT

  Inspired by Driving Miss Daisy,  the 1989 American comedy drama film, I borrowed its title to headline this article. Coincidentally, the f...