Deepfakes: The Surreal World of AI-Powered Video Manipulation on Your iPhone
In just the past few years, a revolutionary technology has emerged that allows anyone to digitally manipulate videos in stunningly realistic ways, right from their smartphone. Known as "deepfakes", this AI-driven capability lets users swap faces, alter speech, and synthesize entire video scenes that appear real – but are completely fabricated. As deepfake apps proliferate for iOS devices, it‘s opening up a bizarre new world of creative expression, as well as some chilling possibilities for deception and abuse.
What exactly are deepfakes? In simple terms, they are videos that have been manipulated using artificial intelligence algorithms to make it look like someone is saying or doing something they didn‘t actually say or do. The "deep" refers to deep learning, the AI technique used to analyze large datasets and train neural networks to perform tasks like facial mapping, voice synthesis, and image blending. The end result is fakery that looks and sounds uncannily real to human eyes and ears.
The technology behind deepfakes has progressed at a breakneck pace. Early examples from just 4-5 years ago were fairly easy to recognize as digital forgeries. But the latest generation of deepfake software has become so good at seamlessly fabricating audio and video that it‘s getting harder and harder to spot the phonies. Some experts predict that within 2-3 years, deepfakes will be virtually indistinguishable from the real thing to the average person.
While the first deepfakes required costly equipment and technical skill to pull off, now anyone can create their own highly convincing fakes with a few taps on their iPhone. A rapidly growing collection of iOS apps have emerged that place the power of deep fakery in the palm of your hand.
Some of the most popular deepfake apps for iPhone include:
Reface: Lets you swap your face onto GIFs, videos and memes. Described as "Photoshop for video."
Wombo: Enables full-body deepfakes, 3D animation, and lip syncing set to music.
DeepArt Effects: Turn your videos into moving paintings/artwork based on art styles you choose.
FaceApp: The aging filter went viral, but it also offers face swapping, gender bending, etc.
Other apps like Morphin, Xpression, MugLife, and Impressions also offer various flavors of face swapping, voice changing, and deepfake creation tools. Apple‘s own Clips app contains a feature called Memoji that uses deepfake-style facial mapping to create 3D cartoon avatars that move based on your facial expressions.
The deepfake apps advertise themselves mainly as fun tools for creative expression, humor, and music/dance videos. Many have sharing integrations with social media platforms like TikTok, Snapchat and Instagram where deepfake memes and face swaps flourish as a form of entertainment.
However, there is also a dark side to the rising accessibility of deepfake technology on iPhones and other smartphones. Malicious actors can misuse these tools to create deceptive media for a variety of unethical and illegal purposes:
Spreading mis/disinformation: Fake videos of politicians and celebrities appearing to say or do outrageous things.
Fraud: Scammers cloning people‘s faces/voices to fool others or defeat authentication safeguards.
Harassment and abuse: Bullies using face swaps for cruel jokes and impersonation of victims.
Revenge p*rn: Placing someone‘s likeness into explicit sexual content without consent.
These unsavory use cases have led to growing concern about how to combat deepfake apps and content designed for deception and exploitation. Tech giants like Apple and Google are working on ways to detect synthesized media and label it to alert users. Blockchain-based authentication is also being explored as an "immune system" against fakes.
Currently, the laws and regulations around deepfakes are fuzzy and inconsistent. A few US states like California, New York and Virginia have passed narrow bans on certain malicious uses of deepfakes, but there are no overarching federal rules as of yet. Legislative proposals are being debated to place restrictions on deepfakes used for election interference, fraud, and nonconsensual p*rn – but it‘s a complex challenge.
The fundamental problem is that deepfake technology itself is simply a set of AI tools that can be used for good or ill. Banning or restricting the tech itself is extremely difficult, especially as it rapidly evolves and becomes more accessible through apps. There are also thorny free speech issues involved with regulating this new medium of synthetic media and expression.
Even as efforts mount to combat malicious deepfakes, there are also promising legitimate applications being explored:
Movies/TV can use deepfakes to more easily create visual effects, de-age actors, or reanimate old footage in new ways.
Educators can bring historical figures to life with deepfaked video lessons and digital avatars.
Businesses can generate personalized AI spokesman, customer service agents, and training videos at scale.
Activists can raise awareness with provocative deepfakes, like swapping male and female actors to highlight gender disparities.
Accessibility tools can automatically animate sign language from speech or dynamically dub videos into other languages.
Many compare deepfakes today to the early days of Photoshop. At first it was a curiosity, then a creative novelty. Over time it became an indispensable tool for digital artists and designers. But it also became a source of fakery, conspiracies, and distorted body images that pervaded society. Like Photoshop, deepfakes are poised to transform creative industries and parts of the digital economy – while simultaneously ushering in a new era of media manipulation.
As an iPhone user, what does the surreal world of deepfakes mean for you? On one hand, apps like Reface and Wombo offer fun new ways to express yourself, make creative videos, and participate in playful memes and trends. On the other hand, we all need to be more aware than ever that video evidence can be falsified, and to think critically about the media we consume and share.
Some advice for navigating the new reality of deepfakes:
Be very skeptical of inflammatory video clips of public figures saying/doing outrageous things, especially if not corroborated by mainstream media. If it seems too crazy to be true, check trusted sources before amplifying.
If someone sends you a video that seems out of character for them or makes claims you‘re unsure about, contact them through another channel to verify it‘s really from them.
Watch out for telltale glitches that can indicate a deepfake: flickering, unnatural facial movements, inconsistent lighting/shadows, blurry patches, misaligned lip syncing, robotic-sounding voices.
Update your iOS devices with the latest security features, like Apple‘s new iMessage Contact Key Verification that can help prove a sender is real.
Support efforts for digital literacy education to help people spot fake media and manipulated content.
If you use face swapping apps and deepfakes for creative purposes, be ethical and avoid using others‘ likenesses without permission in ways they might not approve of. Don‘t assume anonymity protects you from consequences.
Advocate for sensible regulations and platform policies that target harmful uses of deepfakes without squelching benign creative expression.
The genie is out of the bottle with synthetic media and deepfakes. These AI-powered tools are here to stay and will only get more sophisticated over time. Our job is to understand the technology, use it ethically, and adapt our media literacy for a world where seeing is no longer always believing. It‘s a strange new frontier – but with the right knowledge, tools and policies in place, we can harness the surreal power of deepfakes for more good than harm.
