Research & Publications

Explore the scientific foundation behind Earkick's AI-powered mental health technology. Our research spans speech emotion recognition, vocal biomarkers, and AI-driven mental health assessment.

AI Companions in 2025: One Term, Many Realities

K. A. Stephan, H. Bay2025Whitepaper

AI companions now span wellness coaches, romantic partners, general assistants, and proto-clinical advisors. These categories differ in intent, incentives, and risk, yet blurred definitions still cloud safety testing and regulation. When engagement is optimized for stickiness rather than wellbeing, sycophancy and parasocial entanglement grow unchecked and new mental health risks surface under the radar. Clear recommendations show how policymakers, practitioners, and users can draw the right lines and guide AI companionship toward safe, transparent, and evidence-based growth.

Read Full Article →

New Laws, Same Mission: AI for Mental Health in 2025

K. A. Stephan, H. Bay2025Whitepaper

As AI mental health tools become mainstream, states like Illinois and Nevada are introducing laws to define what these tools can and cannot do. This paper breaks down key regulations like HB 1806, explains why they’re emerging now, and shows how privacy-first platforms like Earkick can stay compliant without compromising their mission: accessible, stigma-free support that doesn’t mimic therapy or store user data.

Read Full Article →

How Teens Use Earkick: A Case-Study of AI-Enabled Self-Help

K. A. Stephan, H. Bay2025Whitepaper

A detailed case-study uncovering how teenagers actively use Earkick’s AI mental health app, highlighting real stories and practical impact on daily stress, emotional struggles, and overall well-being.

Read Full Article →

Employee Mental Health in Workplace: Still Declining in 2025

K. A. Stephan, H. Bay2025Whitepaper

A comprehensive study examining how employee mental health is still declining in 2025, despite the availability of AI chatbots, data-rich companies, engagement analytics, and AI-enhanced HR solutions.

Read Full Article →

Ethical AI for Mental Health: Earkick's Perspective in 2025

K. A. Stephan, H. Bay2025Whitepaper

An in-depth look at how Earkick’s ethical approach to mental health tools supports fair access, transparent use of data, and honest conversations about emotional challenges with a online chat option for employees.

Read Full Article →

Analysis of Emotion Annotation Strength Improves Generalization in Speech Emotion Recognition Models

J. Palotti, G. Narula, L. Raheem, H. Bay2023IEEE/CVF Conference on Computer Vision and Pattern Recognition

A groundbreaking report on how better emotion labeling boosts the reliability of speech emotion systems, sharing key findings that help create better chat to talk to tools for future emotion research projects.

Download PDF
Earkick bear

Try Earkick today!

Track and improve your mental health in real time

Appstore download button
Panda's voice is
Powered By Eleven Labs