What We Learned from Our Public API Launch

A few weeks ago, the Reality Defender team and I took the plunge and launched our public API and SDK with a generous free tier offering 50 scans per month. As a deepfake detection company, we knew there was growing demand for our technology, but we weren't prepared for the breadth and diversity of interest we'd see from around the globe.

The response was truly eye-opening, revealing the state of problems related to deepfakes across industries and geographies. What we discovered in the weeks since launch speaks volumes about the current state of both deepfakes and where detection is needed most.

A Truly Global Problem

Perhaps the most striking revelation was the geographic spread of our early adopters. While we expected strong interest from North America and Europe, we were taken aback by the significant uptake across the Asia-Pacific region. Countries like Malaysia, Singapore, Taiwan, and Indonesia showed particularly robust rates of adoption, suggesting that deepfake threats are a pressing concern across diverse economic and technological landscapes. (This should not be too surprising, however, given our work in combatting deepfakes across Singapore and Japan.)

This global distribution tells us something important: deepfake threats don't respect borders or development levels. Whether it's a fintech startup in Southeast Asia or a major bank in Europe, organizations worldwide are recognizing the need to defend against increasingly sophisticated AI-generated content.

Patterns Emerge

Through analyzing how users are implementing our API, three separate patterns have emerged that reflect the current threat landscape.

Identity verification has become a common application across all sectors. Organizations are integrating deepfake detection into their existing KYC processes, adding an extra layer of security to ensure that a person in question is who they claim to be. This makes perfect sense given the rise of synthetic identity fraud and the increasing sophistication of deepfake technology.

Fraud detection is another popular use case, particularly among financial services companies. These organizations are signing up to use our technology to analyze communications and media for signs of manipulation, helping to prevent social engineering attacks that leverage synthetic audio or images to impersonate trusted individuals.

We've also observed interest in broad use of synthetic media detection, which encompasses everything from content moderation to brand protection. This category has seen repeatedly high interest levels, likely driven by the growing awareness of how deepfakes can be weaponized for disinformation campaigns or corporate sabotage.

The variety of organizations implementing our API has been remarkable. Financial services companies are leading the charge, which aligns with their position on the front lines of fraud prevention, while serving as a logical follow-on to existing work we've conducted in the industry. Banks and fintech companies are particularly active, recognizing that their digital-first approach to customer interactions makes them vulnerable to sophisticated impersonation attacks.

Technology companies represent another significant segment of our user base, with major firms exploring how to integrate deepfake detection into their existing products and services. This suggests that we are seeing the beginning of a broader ecosystem where deepfake detection becomes a standard component of digital infrastructure.

Media and journalism organizations have also shown strong interest, reflecting the industry's urgent need to combat disinformation and maintain editorial integrity in an era of synthetic content. Fact-checking organizations are now increasingly leveraging our technology to verify the authenticity of multimedia content in their investigations.

Perhaps most encouraging has been the adoption by academic institutions worldwide. Universities are using our API for research purposes, helping to advance the broader understanding of synthetic media threats and developing new approaches to detection. This academic interest suggests that the next generation of technologists will be better equipped to handle these challenges.

Finally, government contractors and security-focused organizations have also been early adopters, understanding that the implications of deepfake technology extend far beyond commercial fraud to national security and public safety concerns.

What's Next

The diversity and scale of interest in our public API has provided strong validation for our approach to deepfake detection. The fact that organizations across such varied industries and geographies are actively seeking solutions tells us that deepfake threats have moved from theoretical concern to practical business problem.

The global nature of the response reinforces our belief that effective deepfake detection requires continuous innovation and adaptation. Threat actors operate internationally, and their techniques evolve rapidly. Organizations need detection capabilities that can keep pace with this evolving landscape while remaining accessible and easy to integrate.

As always, our team is working hard to update, upgrade, iterate, and improve on our existing models, which will then make their way to our public API and free tier. So too will detection across modalities like video and text, set to be a part of the public API in the coming months.

Yet as we continue to analyze the data from our API launch, one thing is clear: the demand for reliable, real-time deepfake detection spans industries, geographies, and use cases. The challenge now is ensuring that defensive technologies can evolve as quickly as the threats they're designed to counter. The future of digital trust may well depend on how effectively we can democratize access to these defensive technologies while maintaining their effectiveness against increasingly sophisticated attacks.