The digital world is transforming as generative AI advances, leading to a surge in synthetic media, also known as “deepfakes.” Public concern is growing: a Deloitte study found that nearly 70% of people familiar with generative AI worry about synthetic content being used for deception. About 60% struggle to distinguish between AI-created and human-made content, while 84% strongly support clear labelling of AI-generated media.
This challenge coincides with the rapid expansion of the AI market, projected to grow from $189 billion in 2023 to approximately $4.8 trillion by 2033, according to UNCTAD. As AI becomes increasingly crucial in media and finance, reliable tools to authenticate digital content are becoming more necessary. IdentifAI aims to fill this gap.
In a conversation with TFN, Marco Ramilli, co-founder of IdentifAI, explains, “IdentifAI is developing proprietary deepfake detection technology because the demand for reliable media verification tools is surging. With AI expected to grow from 7% to 29% of the frontier technology market over the next decade, this problem will only intensify.”
Following an initial seed round of €2.2 million less than a year ago, IdentifAI has now raised a total of €7.2 million, including a recent €5 million round led by United Ventures. These funds will support IdentifAI’s expansion into key European markets and the United States, as well as ongoing research, development, and refinement of its products.
This latest funding round is perfectly timed, according to Castaldo: “With AI’s exponential growth, we’re working against the clock. The new funding couldn’t come at a better time.”
The story of IdentifAI: From a personal spark to a global mission
IdentifAI’s origin stems from a moment that vividly demonstrated the fragility of visual trust. Ramilli recalls, “Imagine the scene: I’m in a taxi, stuck in traffic, and a photo pops up on my phone: the Pope in an oversized white puffer jacket. At first, I fell for it too. Within minutes, I discovered it was a deepfake. Meanwhile, on X (formerly Twitter), chaos erupted: some were outraged, others laughed, some thought it was real. That moment revealed a fundamental fracture: visual trust, something we’d always taken for granted, was no longer intact.”
As an experienced security researcher and entrepreneur, Ramilli had encountered many internet vulnerabilities, but the implications here were far more profound. “If we can no longer trust our eyes, the entire information ecosystem, from journalism and insurance to banking, begins to crumble. When I surveyed the market, I found no reliable solution to certify the origin of images, video, or audio. Banks require identity verification, media outlets aim to avoid embarrassing mistakes, and insurers need accurate remote damage assessments. There was a massive gap and exploding demand.”
This combination of personal experience and urgent industrial need shaped IdentifAI’s founding mission: “Our idea was born from the gut, developed with the head, and is now driven by a global need to restore transparency using models that can determine, in seconds, whether a human or a machine created content.”
“From day one, our goal has been to build a global company, because truth is a universal value that needs protection worldwide,” Ramilli emphasises. As public demand for trustworthy verification grows, IdentifAI stands ready to serve both individuals and organisations.
IdentifAI’s commitment to diversity is reflected in its pan-European team. “While female representation remains limited today, we strongly believe that cultural and gender diversity are core values and powerful drivers of improvement and ongoing innovation,” Ramilli notes.
How AI with cybersecurity DNA powers real-time detection
IdentifAI’s technological foundation is rooted in cybersecurity. Ramilli explains, “Our founders come from offensive security and incident response teams, so every module is developed under an ‘assume-bypass’ mindset. We built an internal toolchain that lets us launch new detectors immediately when new image, video, or voice generators appear in the wild.”
The company maintains an active research loop. Discoveries are published and peer-reviewed before integration into production, ensuring continuous improvement and quality assurance. Performance testing utilises both open datasets and proprietary benchmarks, including hard-to-find edge cases, to assess whether their models generalise well rather than merely memorising data.
The platform is a fully proprietary, end-to-end integrated system with internally owned and optimised model architectures, orchestration, and monitoring dashboards. Stateless microservices architecture allows IdentifAI to scale seamlessly from dozens to tens of thousands of verification requests per second while maintaining single-digit millisecond latency.
IdentifAI also offers domain-specific models tailored to finance, healthcare, e-commerce, and the public sector. These specialised models provide enhanced accuracy and regulatory compliance for each industry. On the forensic intelligence side, their top-tier detectors can identify not only synthetic content but also detect whether images or frames have been manipulated using GANs or similar methods, providing both confidence scores and tamper maps in real-time.
Compared to competitors like The Hive, Reality Defender, and DuckDuckgoose, IdentifAI distinguishes itself through its unique combination of cybersecurity expertise, real-time adaptability, and comprehensive forensic intelligence.
What’s next for IdentifAI?
Looking forward, IdentifAI plans to advance its technology relentlessly, making it accessible not only to third-party software but also directly to users, enabling everyone to distinguish between real and synthetic content.
“We will keep pushing the cutting edge. At the same time, we will sharpen our EMEA market strategy, mapping key opportunities with the support of our VP and international team,” Ramilli concludes firmly.
