This Deepfake Tool Turning Images Into Livestreams is Topping the GitHub Charts

Deep-Live-Cam, a deepfake tool that turns photos into live video streams, has quickly become the most liked repository for developers to download on GitHub.

Built primarily in the Python scripting language, Deep-Live-Cam is easy to download and modify, and the highly accessible nature of the Python language means that it can potentially be implemented into applications that will work on phones and other communications devices. 

It is yet another reminder of just how rapidly deepfake technology is becoming a security concern, particularly as it pertains to individuals’ and businesses’ rights.

Generative AI fraud is expected to cost the U.S. more than $40 billion annually by 2027, at a growth rate of 32% year-on-year, Deloitte projected in May.

Experts argue that the rising popularity of deepfakes is largely driven by the difficulty traditional detection methods face in identifying them, making them more appealing for use in scams.

It’s relatively easy to ignore a scam email, but a video call from a video feed that looks like a friend, business partner, or relative will be much easier to fall for. 

On the GitHub repository page, there are several requests by other developers to tighten the ethical considerations of the tool.

Lead developer Kenneth Estanislao told Decrypt the tool represents an “interesting use case” mostly for entertainment, interviews, and hiding some scars, or “just plain fun.”

“Yes, it might be used for malicious purposes, but it’s like giving a knife to a person,” Estanislao said. “They do what they can do. I’m hoping that they use it for the above aforementioned or anything useful and not with malicious intent.”

In any case, the software’s GitHub disclaimer notes: “Developers of this software are aware of its possible unethical applications and are committed to taking preventative measures against them.”

In one example, the GitHub repository description features a tutorial and video featuring the likeness of Elon Musk.

Another concern is that once the repository has been downloaded from GitHub, it can be further modified and used by developers with minimal insight into its use.

While the application was first released last year—with its most recent update three weeks ago—it is currently enjoying a surge of popularity, with more than 3,500 “stars” earned today alone, against a total of 20,600 stars in total. 

For context, the next most popular GitHub repository today is a blockchain tool—Shardeum—with just 1,250 stars.

It comes a day after the US Senate passed a bill to outlaw pornographic deepfakes. Broader laws that would protect people from the technology, meanwhile, are only just beginning to be drawn up.

Edited by Sebastian Sinclair

Generally Intelligent Newsletter

A weekly AI journey narrated by Gen, a generative AI model.

Source: https://decrypt.co/244565/this-deepfake-tool-turning-images-into-livestreams-is-topping-the-github-charts