How a late Brazilian Singer Brought Back Through AI Could Change Your Life, and You May Not Know It

 

Alberto Orrico, Alexandre Camargo, Isaac Machado, João Gabriel Vidal, João Gabriel Maniglia, Mariah Levy

 

 

Imagine a world where you can’t trust your own senses. This reality could come true sooner than you think. You might not know it, but you probably came into contact with AI today. From simple grammar checkers to full-fledged language models like Open AI’s Chat GPT and Google’s Bard, the current state of this technology represents only a fraction of its potential.

Picture this: It’s a Sunday afternoon, you’re on the couch and you receive a video of your favorite popstar committing a felony, or the president saying nuclear war is upon us. What would you do in this situation?

Synthetic media tools based on AI can emulate and mirror human voices, writing, visuals, and mannerisms; and they are now widely available to anyone on the internet. The ethical connotations derived from fabricating the image of an existing individual are nuanced, and reality dictates that the ongoing discussion of AI regulation in the White House affects the entire world, not just US citizens and their fundamental rights.

Last month Volkswagen Brazil aired a commercial featuring a virtually recreated version of late singer Elis Regina, and her virtually recreated voice sang a song from another artist. The ad brought about controversy because although the element of nostalgia was pleasing for some, many argued the singer would’ve never agreed to support the company or sing that song.

The late artist is incapable of giving permission to the use of her image. The gray lines of consent get even more blurred when considering her family approved and even partook in the commercial. Hence the need for a set of Universal Laws that would limit the recreation of one’s likeness using artificial intelligence.

If we try to explain this event through an ethical perspective, we end up reaching Kant’s Universal Law. This concept, unlike the utilitarian method, holds the individual as an end in itself. Basically, if we would apply this logic to the example, it would be considered unethical to use the late singer image as a means for profitable ends since she hasn’t given her consent to take part in the ad. The regulation of AI should follow that premise. The reproduction and manipulation of one’s actions should only be allowed when the individual has the capacity to fully consent to it.

The gap between innovation and protection becomes progressively pronounced when analyzing how developing nations are vulnerable to the misuse and ethical nuances of Artificial Intelligences developed by Big Techs located in developed countries.

While the US leads the world in AI-related laws, these are directed at stimulating the economics of an exciting new industry. Regulation on citizen protection has been scarce. This side of the debate has been led by the European Union which has established a set of directives to ethical use of AI and classified its uses into risk levels.

Still, this type of action does little to protect the public, and the current White House debate on AI regulation fails to acknowledge the effect these American AI models have worldwide. With little defense mechanisms, developing countries may not be able to hold either the user or the company accountable for malicious actions.

As the world leader in AI development, it is arguable that the US has a duty to lead future discussions to cooperate on the regulation of this threat on an international level. Taking into consideration that the promotion of Human Rights worldwide is central to the US’s Foreign Policy, consumer countries of the “Global South” should be represented in regulation discussions. More needs to be done to protect not only US citizens, but world citizens as well.

---

Artigo enviado para The Washington Post.