DEEPFAKE VIDEO MAKER

The term 'deepfake' is a combination of 'deep learning' (a subset of AI) and 'fake', reflecting the method used to create the content and its deceptive nature. Advancements in AI technology have allowed for the creation of highly realistic images, videos, and audio that appear to be authentic but are manipulated or entirely fabricated. In video and audio form, deepfakes often involve replacing or altering a person's face, voice, or movements in a way that makes it look like they are saying or doing something they never actually did. While deepfakes are often used for creative, entertainment, and educational purposes, they also raise ethical and legal concerns due to their ability to spread misinformation and violate privacy. Because of this, Kapwing ensures all deepfake content is clearly labeled with a non-removable 'AI-generated' watermark.

Deepfake technology uses advanced AI, particularly deep learning algorithms like Generative Adversarial Networks (GANs), to create highly realistic fake videos, images, or audio. It works by training a model on a large dataset of a person’s face or voice, allowing it to mimic their expressions, movements, and speech. The result is content that looks or sounds like the real person, even though it's entirely fabricated.

No, deepfake videos are not illegal and there is no federal legislation in the United States that bans or regulates deepfakes. However, ten U.S. states have legislation that specifically targets those who create and share explicit deepfake content, including California, where Kapwing is based. Kapwing supports the control and safety of deepfake content, and our engineers work hard to ensure AI content cannot be used for explicit, misleading, or inappropriate themes. See our Terms of Service to read more on how we ensure AI content causes no harm.

In January 2024, Taylor Swift was the victim of a deepfake leak. Fake videos that deceived the general public spread across social media with one video reaching 47 million views on X (Twitter) before it was removed. The incident led to an outcry of US politicians calling for new legislation. Nonetheless, seven months later in August 2024, Donald Trump shared a collection of AI-generated images that falsely depicted Taylor Swift and her fans endorsing his campaign for president. Kapwing's AI research also found that Taylor Swift ranks third among the 10 most deepfaked celebrities on the internet.

It is extremely difficult for the average social media user to detect deepfake content — just 17% of people have completed Kapwing's AI Quiz which requires answering whether celebrity speeches are real or fake. Some of the most obvious giveaways are out-of-sync speech, inconsistent background details, unusual limb or face movements, video glitches, and blurring. Researchers, tech companies, and video experts use frame-by-frame analysis alongside deepfake detection software to help identify deepfake content and keep the internet a safe space. You can also flag harmful deepfake content with the app or website it appears on.

The history of deepfake content can be traced back to the early 1990s when the development of CGI acted as a catalyst for the deepfake journey ahead. However, it wasn't until late 2017 when a Reddit user coined the term 'deepfake', sharing AI-generated videos that replaced normal faces with those of celebrities. This was the moment deepfakes were 'invented'.

Because of the advanced technology needed to create HD deepfake content, you're unlikely to find a high-quality deepfake generator for free. However, using Kapwing, you can generate 30 seconds of free video content to test our Deepfake Video Maker. Select a Persona (such as Donald Trump or Taylor Swift) or upload a video of yourself, and edit the script to change what the avatar says. Export and download, with speech and lips in sync. If you decide you want to create regular content using your new Persona then you can join Kapwing via a subscription package.