Researchers at Dartmouth College have demonstrated that human eye movement patterns serve as highly distinctive identifiers, effectively functioning as cognitive fingerprints shaped by individual conceptual priorities. Published in the Proceedings of the National Academy of Sciences, the study reveals that how people visually navigate unfamiliar environments reflects stable, personality-level preferences that artificial intelligence can accurately map and recognize. Led by senior author Caroline Robertson, the research team enrolled approximately sixty participants who explored everyday scenes through virtual reality headsets equipped with eye-tracking technology. Over sixteen-second intervals, participants freely scanned each environment while the headsets recorded gaze duration and fixation points. The team processed this data using a multi-layered AI framework comprising machine learning models, computer vision systems, and large language models. Rather than analyzing raw visual data alone, the large language models interpreted the thematic and conceptual relationships between fixated objects, generating contextual narratives that captured why specific items drew attention. The results indicated that gaze patterns are significantly more individualized than previously recognized. Participants progressed through three distinct perceptual stages: initial spatial orientation toward the horizon and center, followed by fixation on prominent visual elements, and finally a shift toward interpreting the conceptual meaning of objects. The AI models successfully identified individuals based on these patterns, with the large language model proving most accurate by mapping abstract conceptual connections rather than physical attributes. Notably, these visual preferences remained stable over time. When a subset of participants retested a week later in novel environments, earlier gaze data accurately predicted their new attentional focus. The findings carry significant implications for both data privacy and clinical neuroscience. As virtual and augmented reality systems become increasingly prevalent, the study suggests that passive eye-tracking may inadvertently expose deeply personal information to advertisers and third parties, surpassing current digital tracking capabilities in granularity. Conversely, the stability and consistency of gaze patterns present promising clinical applications. Researchers propose that this methodology could serve as a diagnostic marker for neurodevelopmental conditions, particularly autism spectrum disorder. By differentiating between reduced visual engagement and conceptual processing deficits, the approach may facilitate earlier identification of symptoms, potentially lowering the current national average diagnosis age from four to two years old. Future investigations will explore multimodal systems that integrate visual and cognitive attention metrics, alongside cross-cultural analyses of conceptual priorities. The Dartmouth team's integration of large language models with visual attention tracking establishes a novel analytical paradigm, positioning eye movement analysis as a quantitative tool for understanding human cognition, privacy boundaries, and clinical diagnostics.