Combatting Deep Fake Images and Revenge Porn: New Challenges and Solutions
The Rise of Deep Fakes and Revenge Porn
Thousands of innocent women have fallen victim to the disturbing trend of intimate pictures being circulated online without their consent. This disturbing trend has been further exacerbated by the emergence of new artificial intelligence (AI) applications that have the capability to transform ordinary images from social media into fabricated pornographic videos.
Recently, offensive AI-created images of Taylor Swift went viral on social media platform X (formerly known as Twitter), prompting an unprecedented uproar against deep fakes and revenge porn. Twitter, having downsized its content moderation staff, took swift action by blocking searches for the pop star’s name, as millions of these manipulated images garnered extensive attention.
It’s important to note that the harmful effects of these manipulative AI technologies extend beyond the creation of fake pornographic images. Other deceptive practices include the manipulation of various kinds of images and voices. For instance, during the New Hampshire primary, some voters received robocalls featuring fake recordings of President Biden, urging them not to vote. Additionally, there has been an alarming rise in deep-fake phone calls targeting the elderly, with scammers falsely posing as their relatives.
The Legal Landscape and Legislative Efforts
While some states such as Virginia and California have taken measures to address the threats posed by deep fakes and revenge porn, the effectiveness of these legal bans remains to be seen. A notable example is Massachusetts State Senator Barry Finegold, an advocate for the prevention of deep fakes in political campaign advertising, who has introduced a bill to outlaw their use. However, the Supreme Court’s interpretation, which may consider First Amendment limitations, will ultimately shape the extent of legal protections. Interestingly, in a 2002 case, the Court struck down a ban on computer-generated child pornography.
The Alarming Growth of Deep Fakes
According to a report by the cybersecurity group Home Security Heroes, approximately 100,000 deep-fake videos, 98 percent of which were pornographic in nature, circulated online last year. This staggering number reflects a significant six-fold increase since 2019, highlighting the urgency of finding effective solutions.
Legislative Advocacy and Technological Innovations
Recognizing the urgent need to combat the spread of non-consensual, digitally altered intimate images, Representative Joe Morelle, a New York Democrat, has spearheaded federal legislation to ban their circulation nationwide. Morelle capitalized on the aftermath of the Taylor Swift deep-fake attacks to rally support for his bill, emphasizing that deepfakes affect not only celebrities, but countless women and girls on a daily basis.
Alongside legal approaches, there is a growing reliance on technological solutions. The StopNCII.org website, initially designed to assist in the removal of non-consensual posts online, has garnered support from major social media services such as Facebook, Reddit, and Snap, as well as porn site Pornhub. Through collaboration, these platforms help to eliminate offending images using digital fingerprints to identify and delete them. To date, their efforts have resulted in the removal of over 200,000 images since 2015.
Additionally, the Biden administration is advocating for AI companies to incorporate a digital watermark into all AI-generated content, making it easily identifiable as fake material. While such watermarking may have proven useful in identifying the Taylor Swift fakes, which were created using Microsoft’s Designer app, concerns remain that bad actors may find ways to evade these watermarks, or even exploit their usage on authentic media.
The Continuous Battle for Online Vigilance
Until more comprehensive solutions are developed, social media platforms, alongside their users, must maintain hyper-vigilance against the proliferation of deep fakes and revenge porn. Given the potential harms inflicted on individuals and wider society, ongoing efforts are necessary to safeguard digital spaces and protect victims from this malicious and exploitative form of online harassment.
Aaron Pressman can be reached at [email protected]. Follow him on Twitter @ampressman.
