Using Our Powers For Good: Deepfakes in the Digital Age

Written by Don Lariviere

‍

The digital world of 2024 is becoming increasingly indistinguishable from reality. Deepfakes – those eerily convincing AI-generated imitations – are no longer the stuff of sci-fi thrillers. They're here, lurking in our social media feeds, news cycles, and even personal interactions.

The tech world is scrambling, trying to develop tools to detect these digital phantoms. It's an arms race, with deepfake creators constantly upping their game, making detection an ever-moving target. We're not just protecting our newsfeeds and inboxes, we're safeguarding the truth itself. 

Amidst the technological and legislative battles, there's a growing recognition that the most powerful weapon against deepfakes might be us – the everyday users. We need to be have a healthy skepticism, to question what we see and hear, to fact-check before we share.

Cutting Through The Clutter: Real Intelligence on Deepfakes

But while deepfakes are often associated with malicious intent, there are several positive applications where this technology has been used creatively or beneficially. These positive applications of deepfakes underscore the promise of artificial intelligence!

Healthcare

The company Voicery has used deepfake voice technology to create personalized, synthetic voices for individuals with ALS (Amyotrophic Lateral Sclerosis) who are losing the ability to speak. By analyzing their voice recordings, the technology generates a deepfake voice that closely matches their natural one, allowing them to communicate with a voice that still feels like their own.

DeepMind has worked with patients who have lost facial features due to cancer treatments. By analyzing old photos and videos, technology creates reconstructions that allow patients to visualize themselves as they looked before treatment, which can have therapeutic benefits.

Deepfakes could create personalized medical explanations tailored to individual patients' needs and understanding, like a doctor explaining a complex procedure in a patient's native language, using visuals and metaphors they can easily relate to. They might also be used to create virtual therapists or support groups, offering accessible and affordable mental health care. This could help reduce stigma and increase access to mental health resources, especially in underserved communities.

Education

TIME magazine used deepfake technology a few years ago to create an interactive exhibit called “The March.” This project brought Martin Luther King Jr.’s iconic “I Have a Dream” speech to life, allowing modern audiences to experience his 1963 March on Washington speech in an immersive way. The deepfake provided a powerful educational tool, preserving and enhancing a crucial moment in history.

Deepfake technology has been used to revive and preserve endangered or extinct languages by generating realistic spoken word samples. By analyzing old recordings, technology creates a lifelike simulation of how native speakers once sounded, contributing to the preservation and revitalization of cultural heritage. Museums are also recreating historical figures, allowing visitors to interact with digital versions of people like Albert Einstein or Anne Frank. 

Deepfakes could make education more accessible for students with disabilities, like a deepfake sign language interpreter translating a lecture in real-time, or a deepfake voice reading aloud a textbook for students with visual impairments. 

Entertainment

Deepfake technology was used in the recent Disney+ series The Mandalorian and to recreate a younger version of Luke Skywalker, played by Mark Hamill. The technology allowed the character to appear in new scenes, despite the actor being much older than his character. 

And, it was used to complete scenes for actors who passed away during filming, like Carrie Fisher in Rogue One: A Star Wars Story. Fans appreciated the authenticity and continuity it brought to the beloved Star Wars universe.

Artist Bill Posters created a series of deepfake videos for a project called “Spectre,” where famous figures like Mark Zuckerberg and Kim Kardashian made unsettling statements about data privacy and control. The project was intended to critique the power of tech companies and provoke thought about digital ethics.

But…How Do They Do It?

Many high-quality deepfakes are primarily created using a type of artificial intelligence called Generative Adversarial Networks (GANs). Here’s how they work: 

Two Neural Networks: GANs consist of two neural networks working in tandem: a generator and a discriminator. The generator's job is to create fake images or videos. It's fed a massive dataset of images or videos of the target person and learns to mimic their facial expressions, movements, and even voice.

The discriminator acts as a judge. It's also trained on real images or videos of the target person. It tries to distinguish between the real content and the fakes created by the generator. 

The generator and discriminator are then pitted against each other. The generator tries to fool the discriminator, while the discriminator tries to catch the generator's fakes. Through this adversarial process, both networks improve. The generator gets better at creating more convincing fakes, and the discriminator becomes more adept at spotting them.

Additional Techniques Used in Deepfakes:

• Face-swapping: This technique involves replacing the face of one person in a video with another person's face. It requires precise alignment and blending to achieve a seamless result.
• Facial reenactment: This involves manipulating the facial expressions of a person in a video to match the expressions of another person. It's often used to make someone appear to say or do something they didn't.
• Voice synthesis: This involves using AI to generate a convincing imitation of someone's voice. It can be used in conjunction with face-swapping or facial reenactment to create even more realistic deepfakes. 

Conclusion

Deepfakes, while controversial, present a spectrum of possibilities that extend far beyond entertainment. They offer powerful tools for creative expression and storytelling, educational enrichment, and even historical preservation as we’ve shown here. 

As we navigate the ethical complexities and technological advancements associated with deepfakes, it’s crucial to approach them with a balanced perspective — acknowledging their potential for both positive impact and misuse. Responsible implantation and ongoing dialogue will be key to harnessing the power of deepfakes for a brighter, more innovative future.