Guardrails for AI: How One Startup Keeps Your Face Safe

The last week’s tech gathering in Culver City drew thousands of Hollywood insiders to learn how artificial intelligence can help film. Most talks celebrated the promise of AI, but a conversation with Luke Arrigoni, CEO of Loti AI, offered a different view. Loti scans the web for unapproved deepfakes—videos or photos that copy a person’s face, voice, or likeness—and removes them. Arrigoni explained how his company acts as a safety net for anyone whose image is used without permission, from actors on strike to everyday users who find their photos altered online. Arrigoni’s business has two parts. First, it hunts down illegal or harmful content: the firm uses face‑ and voice‑recognition software to find deepfakes, scams or even nonconsensual nudity, then deletes them. Second, it works with AI developers who want to generate a person’s likeness. Before the model runs, Loti reviews the request and either approves it or blocks it, ensuring that creators can only produce content that is legal and respectful.The company started by tackling “revenge porn” takedowns, a painful problem for many. After the SAG‑AFTRA strike in 2024, Arrigoni pitched the idea to talent agencies. They agreed it was essential, and Loti grew into a broader platform that protects artists’ identities across the internet. When asked about the future, Arrigoni says AI will produce a lot of content but will rarely match the originality of human creators. He worries that deepfakes could become so convincing that people can’t tell real from fake within a few years. He stresses the need for education—starting in kindergarten—to teach critical thinking so people can recognize misinformation. Despite the risks, Arrigoni remains hopeful. He believes that by providing tools and clear rules, his company can keep AI from eroding trust while still allowing creators to innovate. He warns that unchecked optimization could eliminate human judgment, so he works with investors who share his caution. In short, Loti AI is building a digital safety net that lets people own their likenesses and keeps AI from turning the internet into a playground for deepfakes.