AI-generated explicit content and deepfakes represent a troubling form of image-based sexual abuse. This issue came into sharp focus when Taylor Swift became a target, illustrating that such violations could happen to anyone, including your own children.

While I support Taylor Swift, can we expect companies like Open AI and social media platforms such as X (formerly Twitter) to take similar measures to protect ordinary individuals? Unfortunately, the answer appears to be no.

Taylor Swift's case garnered significant attention on social media, with an AI-generated image of her circulating for hours and amassing 45 million views before action was taken. If not for her dedicated fans, the situation could have escalated. However, such deepfake pornography continues to exist, ready to resurface at any moment.

Some might downplay the seriousness of this issue, arguing that it isn't really her. Yet, Taylor Swift did not give consent for her likeness to be used, making it a clear violation. We are living in a digital era where technological advancements challenge our concepts of consent and privacy, and sexual deepfakes represent a dangerous intersection of technology and sexual violence.

Victoria Rousay's insightful thesis delves into this phenomenon, uncovering both the technological capabilities involved and the deeply gendered nature of this abuse. Through her thorough research, she highlights the lasting emotional and psychological damage inflicted on victims, scars that affect not only their digital presence but their very essence.

At the core of this issue is the misuse of digital identities, where innocent images and videos are transformed into instruments of sexual violence. This not only breaches privacy but also assaults personal identity and dignity, predominantly affecting women who find themselves unwillingly featured in AI-generated porn that spreads across the internet and social media.

Rousay's findings reveal the extensive impact of this abuse—physical, emotional, and social—going beyond the initial shock of discovery.

Recent analysis by an independent researcher shared with WIRED shows a staggering increase in nonconsensual deepfake porn videos, with over 244,625 uploaded to the top 35 websites dedicated to this content in the past seven years. The first nine months of this year alone saw 113,000 videos uploaded—a 54% increase from the previous year. Projections indicate that 2023 will surpass the total number of such videos uploaded in all previous years combined.

The Taylor Swift incident exemplifies the dangers of advanced AI technologies and generative tools, leading to the creation of sexual deepfakes as a form of image-based abuse. The conversation must expand beyond individual victims to critique the broader cultural norms that allow sexual violence to flourish, necessitating a serious examination of the ethical and legal challenges ahead.

Unlike traditional deepfakes, which edit real pornographic videos to include celebrities, AI deepfakes use generative algorithms to create sexualized images of public figures and ordinary individuals from minimal prompts. Tools like Microsoft Designer have been exploited for this purpose; however, Microsoft has since taken steps to close these loopholes.

A Microsoft representative stated, “Our Code of Conduct prohibits the creation of adult or non-consensual intimate content. Any repeated attempts to produce such content may lead to loss of access to the service.” They are actively developing measures to improve safety and mitigate misuse of their systems.

The core issue lies in technology that employs sophisticated AI algorithms to manipulate personal images, resulting in non-consensual pornographic material. The alarming aspect is the accessibility of these tools, empowering almost anyone with internet access to create deepfakes.

Creating these deceptive images now requires minimal effort, as a simple online search can lead to AI tools capable of generating explicit content. This democratization of digital deception poses significant risks, as individuals’ consent can be easily overlooked.

As Rousay's research points out, the gendered nature of this abuse is not coincidental; it reflects a societal issue where women’s bodies are commodified, making their digital identities vulnerable to exploitation. The ease of creating and sharing sexual deepfakes starkly highlights the precariousness of digital identity in a landscape where consent can be disregarded with just a few clicks.

This form of abuse is rooted in heteronormative beliefs that prioritize male desires, marginalizing women and LGBTQ+ communities by using their images for non-consensual fantasies.

Let’s pause for a moment to consider the implications of discovering your image manipulated for explicit scenarios circulating online, or falling victim to sextortion. Many victims, including children and young adults, have faced severe consequences, with some tragically taking their own lives due to these experiences.

The psychological impact is profound, with victims grappling with the haunting presence of their digital counterparts engaging in acts they never consented to. The emotional trauma permeates their lives, affecting relationships and mental health.

Moreover, the social stigma attached to being a target of sexual deepfakes can lead to isolation, compounding the harm. Rousay's research demonstrates that the scars of such abuse extend far beyond the digital sphere, leaving lasting impacts on the lives of those affected.

Addressing sexual deepfakes cannot fall solely on the shoulders of victims and communities. Policymakers must act swiftly to establish laws specifically targeting this form of abuse. Technology companies also share the responsibility to create effective systems for identifying and removing non-consensual content from their platforms.

Collective action is crucial. Public-private partnerships can foster advancements in authentication and verification technologies, ensuring the integrity of digital content. By collaborating, we can establish a safer online environment, protecting individuals from the devastating effects of sexual deepfakes.

In conclusion, this issue has been lurking in the shadows, but the recent incident involving Taylor Swift has sparked significant concern. The silver lining is that there is ongoing discourse surrounding this topic, serving as a wake-up call for the companies behind generative AI tools to recognize the need for regulation.

Though these companies may claim their AI lacks sentience and that we are at the nascent stages of artificial intelligence, we must consider the potential harm to society if we fail to act decisively today.

Thank you for your attention.

Sources - Deepfake Porn Is Out of Control — WIRED - What Nobody Is Talking About Is How Generative AI Can Produce Realistic AI Porn — Data Driven Investor