Luke Arrigoni and Loti AI: The Fight Against Deepfakes Intensifies

Last week, the Amazon MGM studios in Culver City hosted the AI on the Lot conference, an event entirely dedicated to artificial intelligence and its impact on cinema. This gathering attracted nearly 2,500 participants, all eager to understand how AI could reshape Hollywood.

Although the panels often avoided directly addressing questions of human creativity, intent, and identity, these themes were at the heart of discussions behind the scenes. Among the most notable exchanges was an interview with Luke Arrigoni, CEO of Loti AI, which particularly caught attention. Loti AI specializes in detecting and removing deepfakes from the internet, whether they involve celebrities or anonymous individuals.

Loti AI's Missions

Arrigoni explains that Loti AI has two main missions. The first is defensive: the company scours the entire web to identify and remove unauthorized deepfakes, including explicit content and scams. Using advanced facial and voice recognition technologies, Loti AI is capable of locating and eliminating these harmful materials. Arrigoni compares this process to a search, detection, and destruction operation.

The second mission is offensive. Loti AI provides creators with the ability to control the use of their image by generative AI companies. These companies must obtain approval from Loti AI before using protected images or works, or risk having their creations removed. Arrigoni emphasizes that creators have the power to set rules and rates for the use of their work.

Origins and Perspectives

Arrigoni began his venture by focusing on the removal of revenge content, a term referring to non-consensual intimate images. The SAG-AFTRA strike accelerated this transition, prompting talent agencies to adopt these services to protect their clients. Having worked at CAA (Creative Artists Agency), Arrigoni reached out to agencies like William Morris Endeavor (WME) to offer his services, and the response was unanimous: "It seems obvious. Let's go."

Regarding the future of AI, Arrigoni remains skeptical about its ability to surpass human creativity. He believes that AI merely predicts an average based on data of varying quality. He mentions that all the data currently examined falls between the 50th and 75th percentile in terms of quality, as the bottom quarter is illiterate and the top quarter does not publish much. According to him, AI lacks a mechanism to reach the 99th percentile, where true exceptional creators reside. Arrigoni also thinks that people will lose their appetite for AI-generated content, and that we will be left with real creators and real people producing art, journalism, or other things that AI cannot replicate.

Upcoming Challenges

Arrigoni predicts that deepfakes will soon become indistinguishable from reality, posing a major challenge for public education. He believes it will likely be impossible to distinguish the true from the false in a year or two. He emphasizes that overcoming this issue will take at least a generation. For him, it is crucial to start teaching critical thinking as early as kindergarten to help individuals navigate a world where the veracity of media is increasingly difficult to discern.

Finally, Arrigoni warns of a future where current technology could dominate, creating a world where mediocrity is pervasive. Such a scenario would, in his view, be disastrous for society.