Keeping Up with Deepfakes

Artificial Intelligence has brought us many wonders, including personalized playlists, smart assistants, and, of course, homework help. But among its more alarming creations is deepfakes, AI-generated images or videos so convincing they can make anyone appear or do anything. From fake celebrity interviews to hoax crime footage, deepfakes blur the line between what's real and what's notl.

As this technology advances faster than ever, the legal system faces an uncomfortable question: How do you prove what's real if AI can fake everything?

The Deepfake Dilemma:

Deepfakes started with simple internet shenanigans, such as swapping one actor's face for another's in a movie scene. However, they have quickly evolved into a serious threat. Today, everyone with a laptop and free software can create highly realistic fake content in minutes.

The danger goes far beyond funny images and memes. Can you imagine a politician admitting to crimes that they never committed? Or a CEO announcing innovations that never happened? These

videos won't just sway elections and markets; they could undermine public trust completely, leaving the public unsure of what, or who, to really believe.

The legal system is built on evidence, and deepfakes attack the very foundation of the phrase "truth in evidence".

Evidence or....... not?

Courts rely heavily on visual proof (security footage, recordings, pictures, etc). But deepfakes make every pixel suspect. How can a judge or jury trust a video when it might be faked or altered? Even experts struggle. Detecting deepfakes often requires sophisticated forensic tools, and the technology that creates them improves faster than the tools that expose them. It’s a digital arms race. One side invents, while the other side investigates.

If a deepfake surfaces in a trial, lawyers must not only prove what is true but also disprove what isn’t. That makes litigation slower, costlier, and, frankly, more confusing than ever.

Who's Reliable when Reality Glitches?

If a deepfake ruins someone’s reputation or causes financial harm, who’s responsible?

• The person who made it?

• The platform that hosted it?

• The AI model that generated it?

So far, the law hasn’t caught up. In the U.S., Section 230 of the Communications Decency Act largely protects platforms from liability for user-generated content. That means if someone uploads a harmful deepfake to social media, the platform might walk free while the victim scrambles for justice.

Some states, like California and Texas, have begun passing deepfake laws, targeting malicious political or non-consensual uses. But enforcing them across borders (or even across the internet) is another story.

Balancing Regulation and Innovation

Governments are racing to regulate deepfakes without stifling innovation in legitimate AI fields like entertainment or education. The European Union’s AI Act proposes labeling requirements for AI-generated content, while U.S. lawmakers are discussing similar “digital watermark” rules.

Still, new laws move slower than new tech. By the time legislation is finalized, the next generation of deepfakes may already look (and sound) indistinguishable from reality.

Deepfakes are making truth seem optional. Let's just hope lawyers don't start citing ChatGPT as witnesses anytime soon.

Sources

Buffett Brief. The Rise of Artificial Intelligence and Deepfakes. July 2023.

Villasenor, John. “Artificial Intelligence, Deepfakes, and the Uncertain Future of Truth.” Brookings, Brookings, 14 Feb. 2019, www.brookings.edu/articles/artificial-intelligence-deepfakes-and-the-uncertain-future-of-truth/.