Deepfakes: The Dark Side of AI that Ruins Women’s Lives

1. Nonconsensual deepfake pornography has exploded in popularity, with women making up 99% of victims.
2. Sophisticated AI now allows amateurs to easily create fake explicit videos and images using real people’s photos.
3. Tech companies profit from the spread of deepfakes but lack incentive to address the problem without regulation.
4. Toxic misogynistic culture in AI research community fuels lack of empathy for female victims.
5. Laws criminalizing deepfakes help but don’t address the root causes enabling their creation.
6. Mental health support urgently needed for survivors along with better tools to remove nonconsensual content.

Taylor Swift is the latest high-profile victim of the disturbing trend of nonconsensual deepfake pornography. This involves taking images of a person from the internet and digitally altering them to create fake explicit content. The videos and images then spread rapidly across social media platforms.

For the victims, discovering deepfakes of themselves can be traumatic. They may worry about the impact on their personal and professional lives if friends, family or colleagues encounter the content. Attempts to remove the images often fail as they continue spreading online. Some even remain for years as the technology does not yet exist to fully erase them.

Taylor Swift’s case highlights the scale of this issue. Over the past year, deepfake detection firm Sensity reported a sixfold increase in nonconsensual deepfake pornography. Their research found women made up 99% of victims.

Part of the rapid growth ties to the increasing sophistication of artificial intelligence (AI) and machine learning. Now an individual can produce a doctored 60-second video using a single image in under 25 minutes. They can do this at no financial cost using free online tools.

Nonconsensual Deepfakes: Why Tech Giants Must Address This Growing Crisis

Experts argue that successfully combatting this abuse means encouraging social media platforms, search engines and financial service providers to stop enabling deepfake creators. All profit either directly or indirectly from the spread of this content, but face little incentive to address the problem.

Britain has passed regulations criminalizing the distribution of deepfakes. Additionally, laws exist holding search engines and user-generated platforms more accountable. The United States lacks similar legal protections, although members of Congress recently introduced a bill allowing victims to sue deepfake creators.

While important progress, experts argue these actions fail to address the root of the problem. Sophie Compton runs an advocacy group against nonconsensual deepfakes. She believes technology companies only respond when profits get impacted. To her, search engines play an instrumental role in reducing access to deepfake imagery and videos.

Professor Hany Farid specializes in digital forensics at UC Berkeley. He echoes this view, arguing that tech firms will continue ignoring the abuse of women for money unless forced to act. In his view, their “moral bankruptcy” ties directly to the toxic, misogynistic culture permeating AI research and development.

There exists a profound lack of empathy toward female victims among the male-dominated teams building these technologies. A recent industry report found that just 28% of tech professionals and 15% of engineers in the United States are women. Anecdotal evidence suggests rampant gender discrimination.

Inside this culture, the non-real nature of deepfakes appears to justify dismissing the very real harm inflicted on victims. More education and awareness-building focused on that harm may prove critical in combatting this crisis. So too could better mental health support systems for survivors along with robust tools for blocking and removing nonconsensual content.

With generative AI advancing rapidly, the window for action narrows. Governments and tech leaders must intervene quickly and decisively. Their inaction risks normalizing the mass exploitation and abuse of women across the internet.

While laws now criminalize distribution of some deepfake content, they do little to deter anonymous abusers who skillfully exploit legal loopholes and hide behind layers of internet infrastructure. The tech giants controlling that infrastructure possess unmatched power to strangle the economic lifelines keeping this industry thriving. Yet within their male-dominated engineering teams, a culture of misogyny prevails. Empathy for victims remains in short supply.

To these teams, deepfakes register as little more than a fascinating technical challenge, divorced from real world consequences. But as generative AI continues advancing exponentially, the window for intervention closes. Inaction risks normalizing the mass exploitation and abuse of women across the internet. If those building the machine lack the conscience to stop it, then lawmakers must intervene to force accountability. Lives hang in the balance.