The “Fake News Hunter”: “From Faces to Voices to Fake Football Players—This Is How We Expose AI Scams”

Castaldo, from Naples, Co-Founder of IdentifAI: “Every Sector Is at Risk”

Hunters of fake news and AI-generated documents are becoming essential bounty killers in a world where nothing can be taken at face value. A quick scroll on social media is enough to realize it. The recent “Gaza resort,” created by AI showing Trump and Netanyahu sunbathing together, is just the latest example. Companies, especially large ones, increasingly need to defend themselves against videos and images that appear entirely real but are in fact fabricated.

“You need the right technology to analyze them. Now even voices can be duplicated, representing the new frontier of AI counterfeiting,” explains Marco Castaldo, originally from Posillipo, Naples, and co-founder of IdentifAI, a start-up that has developed an innovative AI-based platform using “de-generative” intelligence—technology designed to recognize, with high probability, whether an image or video is genuine. “De-generative, in this context, is not a negative word—it’s the opposite of generative. While AI can create, our mission is to uncover.”

Who He Is

Graduated in Economics and Business at Federico II University, Marco Castaldo has extensive experience in finance and understands what companies fear most today. “IdentifAI was founded in July 2024, and by September we had already secured our first contracts. Its verification rate is extremely high—95%, one of the best in the world.”

How Their Work Looks in Practice

“The simplest example is digital onboarding for banks. Even a ten-year-old child can use AI to produce fake documents with photos. Verifying new online customers who want access to financial products and services is of fundamental importance for banks.”

And in other industries? “Insurance, for example. Photos of an accident assessor’s report can be altered to increase damages from one hundred to one thousand euros. The risks are numerous. The FBI has already mapped out 26 different types of AI-related fraud. Misuse can go as far as altering facial identity. Imagine a cybersecurity company working on a project for months—then someone you think is your colleague requests documents and data online. In reality, it’s an AI-generated identity. This has already happened. Large companies are highly sensitive to these risks, and some are already among our clients.”

Even Football Clubs Are at Risk

“AI can even deceive football clubs. Sending scouts around the world is expensive, so instead they receive videos of supposed young talents—dribbling, bicycle kicks, goals in the top corner. All impressive, but all fake. Clubs understandably want to verify whether a video is real before sending someone to evaluate the player in person. Often, these ‘talents’ turn out to be fabricated.”

Fighting Disinformation in the Media

The news industry is also highly vulnerable. “Verification of images and information is critical. Together with major newspapers, we are advancing the ‘Fight for Truth’ project, through which we’ve invited an initial group of journalists to access our platform free of charge as our contribution to combating falsification of reality for malicious purposes. Major national newspapers and TV broadcasters already have contracts with IdentifAI, and we are preparing training programs because knowing whether something is true or false is a right for all citizens.”

The most alarming development, however, is the new frontier of voice forgery—a technologically recent phenomenon. The latest case involved a criminal group tricking several entrepreneurs over the phone using an AI-generated imitation of Defense Minister Guido Crosetto’s voice.

A Clear Business Model

IdentifAI follows a precise business model, targeting two markets: B2B, by selling companies licenses to use IdentifAI technology; and direct interventions when requested by businesses for more targeted operations. “At the moment,” continues Castaldo, “we are working on a major project with the University of Bologna and Federico II University. With Giorgio Ventre, a leading figure in this field, we are developing advanced capabilities for detecting facial manipulations. This is a constant challenge, as artificial intelligence evolves and improves continuously. Fake news hunters must evolve even faster, in order to safeguard the right to truth.”