Watershed
Parsons School of Design: External Engagement Studio Class
Goal
Strategic Deployment of Ethical and Responsive AI for Public Mental Health
Role
Concept Strategist and Research Lead
Tools
Field Research, Physical and Digital Prototypes, Facilitation Tools, Figma
Outcome
Successfully pitched our concept to both clients, with strong encouragement to continue forward
Overview of Clients
Public Policy Lab:
A NYC-based nonprofit that collaborates with governments and nonprofits to improve public services using a human-centered approach, design thinking, and innovative research to address complex social challenges through service and policy making.
Mayor’s Office of Community Mental Health:
NYC Mayor's Office of Community Mental Health is a government office within New York City’s municipal government, working to improve mental health services and support across the city.
Assigned Challenge
Objective
HMW
Criteria
The Public Policy Lab and the Mayor's Office of Community Mental Health challenged us to develop strategies aimed at reversing the worsening mental health trends among young New Yorkers by 2039.
The objective was to explore mental health interventions for young people by examining how policy shapes design and how design, in turn, influences policy outcomes.
How might NYC respond to the evolving mental health needs of young NYers and to the future overall wellbeing of New York City?
To include:
Foresight models to examine trends and speculative scenarios
Domain maps
Environmental scanning
Systems mapping
Journey mapping
Facilitation of workshops
Proposed Solution
Watershed: Harnessing AI to trace pathogenic rhetoric online as a predictor of self or community harm, fostering safer and healthier digital environments. Acting as a clearinghouse for cross-agency collaboration, Watershed delivers rapid, effective interventions by utilizing AI-driven expert recommendations and optimized budget allocations for targeted action.
How it Works:
Data Collection: AI tool detects early indicators of mental health issues through verbal and visual language scans on social media.
2. Passing a Threshold: Once a signal passes a threshold, an automatic alert is sent to relevant experts who assess and validate data collection findings.
3. Alerting Experts: Experts assist in leveraging existing neighborhood interventions and identifying gaps within them. Prioritization of funding allocations are also generated by AI.
4. Resolution: Interventions are appropriately implemented, along with process tracking.
Design Process
Stakeholders: Ecosystem Challenge
A Stakeholder Ecosystem Challenge tool is used to visualize the various actors, non-human agents, and interactions within a system, especially. It helps to identify and understand the relationships between key stakeholders and the devices or systems that mediate these interactions.
Key Components:
Who are the key individuals, groups, or organizations involved in addressing this challenge?
What tools, technologies, or systems are involved in facilitating or mediating interactions between stakeholders?
How do the stakeholders and non-human agents interact with each other within the ecosystem?
Stakeholders: Ecosystem Mapping
A Stakeholder Ecosystem Map was used to document the ecosystem challenge, listing common elements, themes and interactions between stakeholders. By visualizing the relationships and flows within the ecosystem, it provided a clearer understanding of the dynamics at play, revealing opportunities for intervention and improvement in addressing the challenge.
What is the overarching goal or mission of the system, and how does it align with the needs of the stakeholders?
Empathy Map + Domain Map + Assumption Map
Creating an empathy map allowed for a deeper understanding of the needs, emotions, and experiences of the stakeholders involved.
Creating an assumptions map allowed for identifying and visualizing the underlying assumptions that shape decisions and strategies in a project.
Secondary Research: Environmental Scan
As part of our secondary research, 200 mental health signals were gathered from a range of reputable sources, including academic journals, government reports, healthcare organizations, and expert analyses. These signals were carefully analyzed to identify patterns, emerging trends, and key indicators of mental health issues across young NYers.
Creating a domain map allowed for a comprehensive overview of the key concepts, systems, and relationships within the exact problem space.
Interview Insights
Number of Interviews: 8
Who: partner experts, psychologists, therapists
"With younger generations considering technology a necessity, rising tech use is impacting the brain, as neurotoxicity occurs with unregulated and excessive use. The public health message is clear: there needs to be limits, but how those limits should be designed and implemented is currently unknown."
- Andrea Hamilton, Senior Adviser, Behavioral Health, NYC Mayor’s Office of Community Mental Health
Key Insights
We need networks of like-minded individuals to unite and work with legal experts to draft and advocate for the proposed policies, ensuring they are enacted into law
We must ask ourselves: in the realms of addiction and detoxification, what strategies would be most effective to implement on a large scale?
There’s a need for stronger regulation, including monitoring and government intervention, while balancing privacy and personal freedom
Addressing mental health in the digital age requires systemic thinking, clearer language, and proactive policies—before crises occur, not after
Affinity Map
Once all interviews were conducted, plotting an affinity map helped connect individual insights to broader signals, allowing us to identify patterns and trends. These connections were then clustered into wider themes, providing a clearer understanding of the overarching issues and enabling us to draw meaningful conclusions from the data.
Light Green = Interview Insight, Light Purple = Signal, Dark Purple = Theme
Problem Statement
Identified Need
Revised HMW
Define
Heightened exposure to doom-scrolling contributes to the accumulation of neurotoxic "noise," which not only disrupts cognitive functioning but also fosters addictive behaviors, creating a cycle of mental strain and dependency.
We must reinterpret and reframe both the verbal and visual language on social media, not just as a driver of addiction, but as a vital barometer of youth well-being. This shift can guide us toward detoxification, fostering healthier cognitive environments and a more balanced digital culture.
How can we harness the power of language as a biometric tool to proactively safeguard mental health and well-being, fostering a deeper literacy in emotional and cognitive resilience?
The 2x2 matrix was designed to capture the tension between two key axes: Humanism vs Technocentrism and Silence vs Noise, reflecting the dynamics at the intersection of technology, mental health, and well-being. Neurotoxicity is placed in the lower-right quadrant, representing the 2024 landscape where a technocentric approach prioritizes technological growth over mental health. This creates constant "noise” - information overload, overstimulation, and cognitive strain - that fosters neurotoxic effects, leading to addictive behaviors and mental fatigue.
2x2 Scenario Matrix: 2024
By 2039, we envision detoxification moving to the upper-left quadrant, where Humanism and Silence intersect. This shift reflects a more mindful, human-centered approach to technology, where the focus is on cultivating well-being and mental clarity rather than relentless digital consumption. This evolution signifies a transition from reactive tech management to a proactive, human-first approach that safeguards mental health and fosters long-term well-being.
2x2 Scenario Matrix: 2039
Key Challenges to Implementation
-
How might we use data for social good while still protecting individual privacy?
-
How might we capture the expressions of human mental states that AI alone cannot fully measure?
-
How might we overcome resistance to AI emotion analysis and collaboratively integrate this solution into society?
Data Handling and Privacy
Data Encryption
Employing strong encryption techniques to secure user data both during transmission and storage.
Consent and Transparency
Ensuring that users are informed about the data collection process and providing clear options for consent.
Regular Audits
Conducting regular audits of data handling practices to ensure compliance with privacy regulations and identify any potential vulnerabilities.
Privacy by Design
Integrating privacy considerations into the design and development of the AI system from the outset, rather than as an afterthought.
AI Accuracy and Public Trust
The Power of Participatory Design
Amplify voices that go beyond the digital noise of social media.
Break down the stigma surrounding mental health conversations.
Promote deeper understanding of AI’s potential through inclusive design
Reflections
I learned that frequent pivoting, though sometimes challenging, can often lead to the best ideas and solutions. Initially, we may set a direction based on assumptions, but as we gather more data and feedback, the ability to adapt and pivot opens up new possibilities and refines our approach.
Reflecting on the complexities of data and AI, I’ve come to appreciate the intricate balance between innovation and responsibility. While data itself may seem objective, it's influenced by the context in which it is collected, the biases embedded in algorithms, and the interpretations of both humans and machines. AI has immense potential to drive insights and transformation, but its power also brings forth concerns about privacy, fairness, and accountability.
The development of Watershed highlighted the critical balance between AI capabilities and human expertise. I learned that while AI can optimize recommendations and interventions, human judgment and cross-agency cooperation are essential for effective, context-aware action.