# Adam Gross, CEO of HarmonEyes, on AI-Powered Human-State Intelligence

We discussed the Theia model, which analyzes cognitive and emotional states from eye movements.

- **Source:** [Unite.AI](https://www.unite.ai)
- **Published:** 2026-08-27
- **Topic:** [Business](https://brifing.si/c/business)
- **Original article:** <https://www.unite.ai/adam-gross-co-founder-and-ceo-of-harmoneyes-interview-series/>
- **On Brifing:** https://brifing.si/s/f492d8f1-de26-4c16-a0d3-3f743cda1b03

## Introduction: Adam Gross and HarmonEyes

Adam Gross is a serial entrepreneur with nearly 30 years of experience in technology-driven ventures. In 2013, he founded RightEye, an eye-tracking technology used in healthcare, sports, and military applications. In 2024, he launched HarmonEyes to extend this expertise into AI-powered human-state intelligence. His previous companies, KeyClaims and Trending123.com, were acquired. HarmonEyes is developing a foundation model called Theia.

## Theia's Technology

Theia is an AI model that works on ordinary camera-equipped devices without requiring specialized hardware. It measures signals like cognitive load, attention, fatigue, and stress from eye movements. The model is trained on 585 billion eye-movement data points from 16.3 million records. Data is deleted after processing, with no personal information stored. It has applications in automotive, healthcare, education, and defense.

## Age Estimation Technology

HarmonEyes can estimate age ranges from eye movements. Research with 45,696 people showed 94.67% accuracy across 12 age groups. The system analyzes real-time eye movements rather than profiling behavior. Data is deleted after processing. Liveness detection prevents spoofing attempts, ensuring system integrity.

## Technical Process

Theia's age estimation runs entirely on the user's device. The camera calculates gaze vectors, which the SDK converts into features like eye velocity and fixation points. Machine learning algorithms process these to determine age range. All computations occur on-device, with no raw data retention.

## Accuracy and Validation

Theia achieves 94.67% accuracy across 12 age groups, though the study excluded people with vision disorders or neurological conditions. Additional validation is needed for different populations. HarmonEyes tests across ethnicities, glasses wearers, and environmental conditions to ensure broad applicability.

## Ethics and Privacy

HarmonEyes analyzes momentary eye movements rather than profiling behavior or preferences. Data is deleted post-processing, with no personal information collected. This approach ensures users aren't being tracked. It also provides parents assurance about children's safety without invasive monitoring.

## Application Areas

Theia has potential uses in AI assistants, automotive systems, extended reality, defense, healthcare, education, industrial safety, gaming, and market research. For example, it could detect driver fatigue or measure student attention levels. The range of possible applications is extensive.

## Future Plans

HarmonEyes continues improving Theia's accuracy through testing across demographics and environments. The company aims to expand into more industries, particularly seeing significant potential in education and healthcare. Ongoing research focuses on enhancing the model's capabilities and reliability.

## Conclusion and Significance

HarmonEyes' Theia model offers innovative technology for human-state intelligence through eye movement analysis. Its privacy-focused approach and wide applicability make it stand out. AI-powered human-state intelligence could enhance efficiency and safety across multiple sectors, with HarmonEyes positioned as a pioneer in this emerging field.

## Sources covering this story

- https://www.unite.ai/adam-gross-co-founder-and-ceo-of-harmoneyes-interview-series/ — 

---

_Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._
