Superpower Glass raises Vineland socialization by 4.58 points

AI system

System: Superpower Glass: Google Glass with a real-time machine-learning facial emotion classifier and companion smartphone app
Model: Not assessed
Release: Not assessed
Application developer: Stanford University (Wall lab, Division of Systems Medicine and Department of Computer Science)
Role: Not assessed
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Workflow: Real-time emotion classification on a head-worn display. Structured 20-minute home sessions several times a week. Concurrent applied behavioral analysis therapy.
Human oversight: Not assessed
Relation to AI: The emotion classifier, trained on public facial-expression databases and tuned for home conditions, generates the real-time cues the intervention relies on. The Guess the Emotion game within the package does not use the classifier. So part of the effect may come from structured family interaction rather than the AI.
AI contribution: The emotion classifier, trained on public facial-expression databases and tuned for home conditions, generates the real-time cues the intervention relies on. The Guess the Emotion game within the package does not use the classifier. So part of the effect may come from structured family interaction rather than the AI.

Benefits

Outcome identifiers: Not assessed
Benefit: gained 4.58 points more on the Vineland socialization subscale than controls (P = .005)
Category: Improved Learning
Status: Not assessed
Beneficiary groups: People. Society.
Beneficiaries: Not assessed
Reach: Not assessed
Distribution and access: Not assessed
Duration: Not assessed
Expected?: Yes
Costs and harms: Not assessed

Measurement

Studies: Not assessed
Study design: Not assessed
Measurement: Not assessed
Setting and sample: Not assessed
Measures: gained 4.58 points more on the Vineland socialization subscale than controls (P = .005)
Comparison: applied behavioral analysis therapy alone (treatment as usual)
Results: Result: mean [SD] treatment impact, 4.58 [1.62]; P = .005.
Additional value from AI: Not assessed
Limitations: Not assessed
Denominators: Not assessed
Materials for replication: Not assessed

Money

Kind of value: Not assessed
Amount: Not assessed
Currency and prices: Not assessed
Period and unit: Not assessed
Realization: Not assessed
Recipients and payers: Not assessed
Gross or net: Not assessed
Included costs: Not assessed
Comparison: Not assessed
Method: Not assessed
Return and payback: Not assessed

People and places

Participants: Developer: Stanford University (Wall lab, Division of Systems Medicine and Department of Computer Science). Operator: Participating families at home under the Stanford studies.
Jurisdictions: Benefit: Palo Alto, United States. Developer: United States.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Follow-up

Conditions for replication: Child tolerates wearing Google Glass. Family adherence to the session schedule. Classifier accuracy under home lighting and pose variation.
Replication attempts: 1 follow-up entries. Benefit recurred: 1.
Later findings: 1 May 2019. The feasibility cohort was followed by a 71-child randomized clinical trial at Stanford that found a socialization gain comparable with standard-of-care therapy.
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Evidence

Source numbers: AIBD-0058-source-0001–0003
Citations: Numbered references below.
Language: Not assessed
Source relationships: Developer
Original evidence: Not assessed
Access and versions: Not assessed
Supported findings: Not assessed
Evidence links: Not assessed
Evidence strength: Not assessed
Missing evidence: Not assessed

Grades

Assessed benefits: People. Society.
Depth: Not assessed
Reasons: People: The gains are group means (SRS-2 and Vineland), with no individually attributed outcome. Society: Trial evidence of efficacy comparable to standard care, not a measured societal change.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: The makers' claim

Entry details

Kind

Not assessed

Related entries

Both depend on a real-time facial emotion classifier labeling expressions for a user; Face++ and Microsoft's emotion models misread faces systematically, scoring Black people's smiling or ambiguous photos as negative.

Field

Autism intervention / digital therapeutics. Children with autism spectrum disorder aged 6-12.

Purpose

Teach emotion recognition and facial engagement to children with autism through real-time wearable feedback

Dates

Date basis unassessed: 2016 (precision unassessed). Updated: 21 October 2026.

Stage

Not assessed

Version

SHA-256: de130b277dad…

References

Other identifiers: AI Incident Database: 454.
Category mappings: Not assessed
Brief: Not assessed
Revisions: 6 revisions. Latest: 21 October 2026.

Sources

Peer-reviewed study Developer

Voss C, Schwartz J, Daniels J, et al. (2019). Effect of Wearable Digital Intervention for Improving Socialization in Children With Autism Spectrum Disorder: A Randomized Clinical Trial. JAMA Pediatrics 173(5):446-454.
Peer-reviewed study Developer

Daniels J, Schwartz JN, Voss C, et al. (2018). Exploratory study examining the at-home feasibility of a wearable tool for social-affective learning in children with autism. npj Digital Medicine 1:32.