Welcome to the latest session of ExtraMile by YourTechDiet, where we bring you exclusive insights from industry leaders who are leading the way in technology and innovation.
Today, we are having a conversation with Luke Arrigoni, CEO and Founder of Loti AI Inc., a technology in facial recognition and deepfake detection that protects the digital identities of individuals.
Luke’s journey started from working with celebrities at WME to building AI tools that safeguard people’s online presence. He started a side project with his wife, Rebekah, in 2022 that quickly turned into a mission-critical solution, especially after Hollywood’s AI crisis during the 2023 strikes.
In the discussions, Luke shares how Loti’s AI detects and removes deepfakes faster than social media giants, why digital identity protection isn’t just for celebrities but for everyone, the gaps in AI security most of the founders miss, and what’s coming up next for the company after their $16M funding round.
Whether you are a tech enthusiast, a concerned internet user, or a founder in the AI space, these insights are must-reads. Let’s explore.
Hello Luke, Greetings for the day!
1. From working with celebrities at WME to founding Loti, what made you to shift from entertainment AI to digital identity protection?
Luke. The underlying vision behind Loti AI has always been rooted in protecting identity. Back in 2022, my wife Rebekah and I started it as a small side project to help people find and remove unauthorized uses of their likeness online.
Then in 2023, the Hollywood strikes hit, and AI suddenly became a very real and very urgent issue in entertainment. I started reaching out to friends in the industry, asking, “Hey, would this be useful to you?” and the response was overwhelming. The need was immediate and visible, especially for public figures whose images and voices were being used without their consent. What started as a one-off tool quickly became something much bigger. We realized this wasn’t just a Hollywood problem, this is a universal one.
2. Your platform can detect and remove deepfakes within hours, something even social media giants struggle with. How does your AI work differently?
Luke. We’ve built our system with automation and individual identity protection at the core. If a client needs over a thousand takedowns in a single day, there’s no way a human, or even a team of humans, can do that fast enough. You need automation to move at the speed of the internet.
Our system continuously ingests the crawlable parts of the internet, including social media platforms, to detect unauthorized or AI-generated uses of someone’s likeness. Behind the scenes, Loti uses 40 to 50 different machine learning models focused on synthetic media detection, paired with face recognition. The system is effective even when faces are partially covered or altered, just a few reference photos are enough for us to build an accurate match.
Once detected, we automatically issue takedown requests using copyright law, right of publicity protections, and the Tennessee’s ELVIS Act. That combination of real-time detection and automated action is what allows us to do in hours what would normally take days or never happen at all.
3. You have partnered with agencies like WME and now offer consumer protection. How do you balance protecting A-list clients while scaling for everyday users who are concerned about their online presence?
Luke. It really comes down to how we’ve architected the platform. Both our consumer and celebrity plans run on the same core technology and benefit from the same always-on scanning infrastructure. The system doesn’t slow down based on the number of users or queries, it was built to scale from the start.
We do offer more comprehensive protection plans for public figures, who typically face higher volumes of impersonation or targeted abuse. But we’ve made it a priority to ensure that everyday users can access meaningful protection without compromising the speed or quality of our service.
4. The No Fakes Act is bringing attention to digital identity rights. How do you see regulation shaping the future of digital identity between protection and censorship in this new landscape?
Luke. It’s encouraging to see legislation like the NO FAKES Act recognize digital identity rights. It shows policymakers are acknowledging how urgent and personal this issue is. But regulation is always a balancing act. Move too fast, and you risk stifling innovation. Do nothing, and you leave people, especially those without legal resources, unprotected in a rapidly changing digital world.
The internet requires coordination between governments, platforms, and users. Effective regulation has to reflect that complexity. Platforms want to stay open and engaging, but also safe. Trust and Safety teams are under enormous pressure to act quickly and decisively, yet they often lack the tools or scalability to meet that challenge effectively. That’s where we help bridge the gap with automated systems.
People deserve control over how their image and voice are used. That means legislation that isn’t just built for public figures, but that is accessible, enforceable, and forward-looking. If we get that balance right, thoughtful regulation can help create the foundation for a more trustworthy, inclusive, and human-centered internet.
5. With $16M in recent funding, what is next for Loti? Are you looking to expand into new areas like voice cloning protection or corporate reputation management?
Luke. There is an excellent opportunity for Loti AI in both voice recognition and corporate IP protection. The new funding allows us to expand into those areas but also allows us to hire top talent to make our current execution more efficient, open new opportunities but also for us to continue to innovate in our space.
6. After building AI companies for a decade, what surprised you most about how deepfake technology has evolved, what are the risks and possibilities?
Luke. The first original “deep fakes” were these clever GAN systems that were trained for days or weeks to recreate a single kind of image. I had a friend that spent months training a GAN to create pictures of Jellyfish, that’s how specific it was. Now the technology is so generalized it can create images of anything with little to no input data at inference time. It’s incredible to see how far that industry has come.
7. If you were not safeguarding digital identities, what problem would you like to tackle with AI and why?
Luke. I think there is a huge gap in how we consume technology and what we are trying to get out of it. We have named an entire industry “social media,” but often it drives more antisocial behavior than it does bring people together. I think AI could be used to argue our social capabilities to help people identify who in their world needs a hello, who is stressed, who is happy or even encourage personal growth through deeper introspection.
Right now, technology is really focused on creating a better echo chamber because that sells, but I think a compelling use case for AI would be to push people for long term happiness by encouraging small steps today for better social interactions in the future. A sophisticated system could manage those steps.
8. For young founders looking to build in the AI security space, what is one lesson from your journey you wish you would learn earlier?
Luke. Unless you come from a security background you probably are wrong about your assumptions on where the real threats are. We learned quickly how our assumed value proposition was wrong, that really customers wanted something different and we were lucky the same technology solved for that, as well.
Check Out Our Other Informative Interviews:
