When one of my friends sent a picture of our friend group to me a couple weeks ago, I looked at the photo, thought it was cute, and moved on. That is until my friend mentioned that she used AI to input two of our friends into the picture. Once I looked closer, I could see some details in our expressions and the lighting were slightly off, but had she not told me, I don’t think I would have ever second guessed whether the photo was real.
This instance stuck out to me. Not because my friend had done anything wrong, but because it made me realize how quickly AI has become capable of creating content that looks completely authentic. And at first, it seemed like a great tool. Here was an image that captured all my favorite people, even though some of them weren’t able to be present for the original moment. Then, when I later saw a 6ABC Action News article about a woman who was arrested and put in jail when her ex-boyfriend showed law enforcement screenshots of texts “proving” she had violated conditions of her bond, only to be released a couple days later when it was discovered the screenshots were fabricated, I started to consider what the larger implications of this technology could be. If I couldn’t recognize that a harmless photo had been altered, and law enforcement couldn’t tell that a piece of evidence had no real credibility, then we are entering a world where seeing is no longer enough to establish the truth.
For crime victim advocates, this consideration seems especially pertinent. Firstly, generative AI can facilitate abuse through the creation of fake text conversations, cloned voices, and altered images that can be used to harass, intimidate, manipulate, or damage a victim’s image. Even though policies like the Take It Down Act, which requires platforms to remove nonconsensual content upon request, are being implemented to protect against harmful content being spread online, advocates have an important role to play in the way they discuss digital safety with victims. As the National Network to End Domestic Violence’s Safety Net Project explains, by having conversations about online security, it is possible to identify risks that may go unnoticed otherwise. Consider asking:
- Has the person harming you ever had access to your phone, computer, or online accounts?
- Have you received messages, images, or recordings that seem manipulated or created to intimidate you?
- Are there online accounts or devices that you believe are no longer private?
Not only can AI be used as a tool of coercive control, but the technology also presents a barrier to victims being believed. As cases of AI altered evidence become more widespread, I wonder if victims will begin to hear the reaction “How do we know this wasn’t generated?” when they come forward with evidence. While it’s important to keep in mind that digital evidence may require additional examination, that caution should not become an obstacle for a victim seeking help. No matter what capabilities AI has, victims still need a person to listen to them, validate their experiences, and guide them. Generative AI will continue to evolve, meaning conversations about and understanding of technology-facilitated abuse and digital safety must evolve as well.