1.
Bridging the Gap Between Artificial Intelligence
and Natural Connections
Ranak B. Trivedi, PhD
Assistant Professor, Department of Psychiatry and Behavioral Sciences, Stanford University
Core Investigator, Center for Innovation to Implementation, VA Palo Alto
March 24 2020
@ranaktrivedi

2.
Outline
• The Need.
• The Promise.
• The Gaps.
• The Recommendations.

3.
The Need.
• Worldwide, 15% of people are differently abled
• Largest predictor of disability is age
• 16M individuals are living with cognitive disability in the US
• By 2050, the percentage of the world’s population 60 y+ will be 2B
• Familysize is declining
• 2010: 7:1 informal caregivers: 80 y+
• 2050: 3:1 informal caregivers 80 y+

4.
Goals AI Method Examples
Monitoring Health, Assessing
Risk
Machine learning, NLP, speech recognition Wearables (e.g.,Apple Watch), Smart
Clothing
Managing Physical and
Mental Health
Conversational AI, NLP, speechrecognition,
chatbots
Omada Health, Livingo
Coordinating Care Robotics, Conversational AI Voice Activated Digital Assistants, EllieQ
Assisting with ADL Robotics Robear
Companionship Conversational AI, Social Assistive Robotics Voice Activated Digital Assistants (e.g.,
Amazon Echo),Paro
Monitoring Falls Biometric remote monitoring Biotricity
The Promise: AI tools are transforming care

5.
Gap 1: Limited Attention to Social Process
• Social support improves longevity, health, and mitigates adverse effects of chronic and serious
conditions.
• Two types of social support: Emotional, Instrumental
• Current status
• Wearables focus on individual change
• Simplistic focus e.g., gamification and virality
• Absence of principles from socialtheories

6.
Gap 2: Limited Focus on Informal Caregivers
• In USA, 50M adults provide $470 billion worth of unpaid services
• Average of 18 hours/week
• Limited attention to the needs of caregivers
• Remote monitoring provides actionable information to caregivers, but not “how” or emotional
support
$398.1 B $345.4 B $321.4 B

7.
The Recommendations.
1) Harness the power of social processes
• Motivation: Intrinsic vs Extrinsic
• Social contagion: Behaviors and therefore behavior change can spread
2) Rely on behavior and socialtheory:
• Technology-enabled caregivingat home (TECH)
• Dyadic health behavior change model
3) Embed culturalconsiderations in design
• Engage diverse teams and stakeholders at all stages of development
• Engage caregivers from different cultural backgrounds

8.
The Gerontologist, Volume 60, Issue Supplement_1, March 2020, Pages S41–S49, https://doi.org/10.1093/geront/gnz178
The content of this slide may be subject to copyright: please see the slide notes for details.
Technology-enabled caregiving in the home (TECH).

9.
Patient
Stress
Depression
Self-efficacy
Social Support
Functional Limitations
Stress Coping
Caregiver
Stress
Depression
Self-efficacy
Caregiver Burden
Social Support
Stress Coping
Self-management
Stress Management
Depressive Symptoms
Medication Adherence
Symptom Management
Diet
Exercise
Relationship
Dyadic Coping
Communication
Reciprocity
Collaboration
Biomarkers
Quality of life
Hospitalizations
Mortality
Cost
Dyadic Health Behavior Change Model (Trivedi
et al 2016)

10.
Summary
• All individuals, but especially those who are differently abled and dependent, rely on social
connections
• AI should harness the power of these connections into applications
• AI should do this with sensitivity to culturaldifferences

11.
A heartfelt smile, a gentle tone,
A thoughtful word, a tender touch,
A passing act of kindness done-
‘Tis all, but is much.
These are not things that win applause,
No earthly fame awaiteth such,
But surely by the heavenly laws,
They are accounted much.
-Anonymous

12.
Thank you
Ranak Trivedi, PhD
ranakt@Stanford.edu
https://framilyveteranhealth.sites.stanford.edu
@ranaktrivedi