Nuanced, Inc.

Nuanced detects AI-generated images. We help companies like dating apps, news sites, and ad platforms to distinguish between AI-generated and human-authored content.

With AI-generated content on the rise, many contexts require differentiating it from human-authored materials to ensure authenticity and trust in the digital world. Making this distinction is especially crucial when there are possible implications for misinformation, spam, abuse, or cybercrime. Existing technologies are ill-equipped to do so, costing an annual $78 billion spent on handling deepfakes.

Our API allows integrators to use AI to get ahead of AI easily, without having to spin up their own infrastructure. We train custom models in a zero PII environment, removing the risk of compromising user privacy.

Ayman and Dylan met as undergraduate engineering students at the University of Waterloo. Each cofounder has over a decade of engineering and research experience.

Ayman most recently spent 7 years at GitHub, contributing to several areas of the product, including building and scaling abuse mitigation infrastructure for over 100 million users. While at GitHub, she saw first-hand how AI-driven threats were evolving, including witnessing established methods fail to detect user profiles that were generated by ChatGPT. Ayman’s experience at GitHub and her dissertation research in machine learning-based cloud platform security at the University of Oxford drove her to found Nuanced.

Dylan has over a decade of machine learning experience spanning industry and academia, culminating in a notable 6-year tenure as an Applied AI Scientist at Amazon.

They both quickly agreed that online content creation and consumption were rapidly changing and wanted to build something to help online platforms address evolving authenticity concerns. This is why Nuanced provides a user-friendly API that helps restore the authenticity and trust of services.