Social Intelligence & AI Alignment

Understanding how the same abilities that enable cooperation also enable manipulation

Seeing Yourself Seen: Gaze Perception and the Architecture of Social Awareness

Accurately perceiving when one is the focus of someone else's attention is foundational to social cognition. This ability enables empathy and communication, but also creates vulnerability to misinterpretation and manipulation—in both humans and AI systems.

Experience Gaze Perception

Look at the eyes below. Do you feel they're looking at you? This automatic response demonstrates how fundamental gaze perception is to social cognition.

Direct gaze detected

Building Blocks of Social Awareness: From Gaze to Mind

Behavioral and Neural Building Blocks 📄

"Inside the Social Brain: Neural Pathways Illuminating How We Perceive Attention"

🧠 dmPFC, STS, and Insula form core network for self-referential gaze processing
🔗 Gaze perception acts as foundational mechanism for broader social cognition
dmPFC STS Insula

Click to download paper

Computational Mechanisms of Self-Referential Biases

"Slow and Steady Perceives the Gaze: Unveiling Cognitive Biases in Social Judgments"

Less efficient evidence accumulation in schizophrenia
🎯 Perceptual biases toward self-referential responses
⚖️ Compensatory cautiousness maintains function at cost of speed
Drift-Diffusion Model Simulator

Adjust parameters to see how cognitive processes influence gaze perception decisions:

1.0
0.0
1.0
Decision Time: --
Choice: --
Neural Dynamics & Social Functioning 📄

"Neural Conversations Gone Awry: How Brain Connectivity Shapes Gaze Interpretation"

🔄 Disrupted connectivity in schizophrenia during gaze processing
⬇️ Altered top-down inhibitory control from frontal to perceptual regions
📊 Connectivity patterns relate to symptom severity
Information Flow During Gaze Processing

Click to download paper

Bridging Psychology & AI Safety

From human apophenia to AI hallucinations, from social cognition to alignment—my interdisciplinary approach offers unique insights for building safer, more predictable AI systems.

Explore My Full Research