AI Ethics & Compliance

AI Code of Conduct

Our commitment to responsible AI development and deployment in healthcare service coordination.

Our AI Ethics Principles

These principles guide every aspect of our AI development, from initial design through deployment and ongoing maintenance.

Human-Centered Design

Our AI systems are designed to augment human decision-making, not replace it. Healthcare professionals maintain ultimate responsibility for participant care decisions.

AI provides insights, humans make care decisions
Clear indicators when AI is being used
Easy override of AI recommendations
Human review required for critical decisions

Privacy by Design

Participant privacy is fundamental to our AI development. All AI processing happens within our dedicated secure GCP environment with zero external AI service calls.

Private AI processing only - no external AI services
No participant data shared with AI vendors
Encrypted data at rest and in transit
Automatic data anonymization for AI training

Transparency & Explainability

Our AI systems provide clear explanations for their recommendations, enabling healthcare professionals to understand and validate AI insights.

Plain-language explanations of AI decisions
Confidence scores for all AI predictions
Source data transparency for recommendations
Regular AI performance reporting

Fairness & Non-Discrimination

Our AI models are regularly audited for bias and trained to provide equitable care recommendations across all participant demographics.

Regular bias testing across demographic groups
Diverse training data representation
Equal accuracy across participant populations
Ongoing monitoring for disparate impact

Safety & Reliability

Our AI systems undergo rigorous testing and include multiple safeguards to ensure participant safety and system reliability.

Extensive pre-deployment testing
Continuous monitoring of AI performance
Automatic alerts for unusual patterns
Graceful degradation if AI systems fail

Continuous Improvement

We continuously update our AI systems based on healthcare professional feedback, regulatory changes, and advances in AI safety research.

Regular model updates and improvements
User feedback integration
Compliance with evolving regulations
Investment in AI safety research

Security & Compliance Controls

Comprehensive security controls implementing industry best practices for healthcare data protection.

Infrastructure Security

IS-1
Implemented

Unique Account Authentication Enforced

Authentication to systems and applications requires unique username and password or authorized SSH keys

IS-2
Implemented

Production Application Access Restricted

System access restricted to authorized access only

IS-3
Implemented

Production Database Access Restricted

Privileged access to databases restricted to authorized users with business need

IS-4
Implemented

Firewall Access Restricted

Privileged access to firewall restricted to authorized users with business need

IS-5
Implemented

Unique Network System Authentication Enforced

Authentication to production network requires unique usernames and passwords or authorized SSH keys

IS-6
Implemented

Remote Access Encrypted Enforced

Production systems can only be remotely accessed by authorized employees via approved encrypted connection

IS-7
Implemented

Network Firewalls Utilized

Firewalls are used and configured to prevent unauthorized access

IS-8
Implemented

Network and System Hardening Standards Maintained

Network and system hardening standards are documented, based on industry best practices, and reviewed annually

Organizational Security

OS-1
Implemented

Asset Disposal Procedures Utilized

Electronic media containing confidential information is purged or destroyed per best practices with certificates of destruction

OS-2
Implemented

Portable Media Encrypted

Portable and removable media devices are encrypted when used

OS-3
Implemented

Employee Background Checks Performed

Background checks are performed on new employees

OS-4
Implemented

Performance Evaluations Conducted

Managers complete performance evaluations for direct reports at least annually

OS-5
Implemented

Password Policy Enforced

Passwords for in-scope system components are configured according to company policy

OS-6
Implemented

Visitor Procedures Enforced

Visitors must sign-in, wear visitor badge, and be escorted by authorized employee when accessing secure areas

OS-7
Implemented

Code of Conduct Enforced

Formalized Code of Conduct demonstrates importance of integrity and ethical values, included in Employee Handbook

OS-8
Implemented

Code of Conduct Acknowledged by New Employees

New employees sign statement signifying receipt, understanding, and agreement to follow Code of Conduct and policies

Product Security

PS-1
Implemented

Data Encryption Utilized

Datastores housing sensitive customer data are encrypted at rest

PS-2
Implemented

Control Self-Assessments Conducted

Control self-assessments performed at least annually with corrective actions based on findings

PS-3
Implemented

Penetration Testing Performed

Penetration testing performed at least annually with remediation plan and vulnerability fixes per SLAs

PS-4
Implemented

Data Transmission Encrypted

Secure data transmission protocols encrypt confidential and sensitive data when transmitted over public networks

PS-5
Implemented

Vulnerability and System Monitoring Procedures Established

Formal policies outline requirements for vulnerability management and system monitoring

Internal Security Procedures

IP-1
Implemented

Continuity and Disaster Recovery Plans Established

Business Continuity and Disaster Recovery Plans outline communication plans to maintain information security continuity

IP-2
Implemented

Continuity and Disaster Recovery Plans Tested

Documented business continuity/disaster recovery (BC/DR) plan is tested at least annually

IP-3
Implemented

Cybersecurity Insurance Maintained

Cybersecurity insurance maintained to mitigate financial impact of business disruptions

IP-4
Implemented

Configuration Management System Established

Configuration management procedure ensures system configurations are deployed consistently

IP-5
Implemented

Development Lifecycle Established

Formal systems development life cycle (SDLC) methodology governs development, acquisition, implementation, and maintenance

IP-6
Implemented

Board Oversight Briefings Conducted

Board of directors briefed by senior management at least annually on cybersecurity and privacy risk

IP-7
Implemented

Security Policies Established and Reviewed

Information security policies and procedures are documented and reviewed at least annually

IP-8
Implemented

Incident Response Plan Tested

Incident response plan tested at least annually

IP-9
Implemented

Risk Assessments Performed

Risk assessments performed at least annually with threat identification and formal risk assessment including fraud consideration

Data and Privacy

DP-1
Implemented

Data Retention Procedures Established

Formal retention and disposal procedures guide secure retention and disposal of company and customer data

DP-2
Implemented

Customer Data Deleted Upon Leaving

Customer data containing confidential information is purged or removed from application environment per best practices when customers leave

DP-3
Implemented

Data Classification Policy Established

Data classification policy ensures confidential data is properly secured and restricted to authorized personnel

AI Governance

AI-1
Implemented

Private AI Processing Only

All AI operations execute within our dedicated secure GCP environment - zero external AI service calls to OpenAI, Anthropic, etc.

AI-2
Implemented

AI Model Validation and Testing

Rigorous pre-deployment testing including accuracy, safety, and bias validation across demographic groups

AI-3
Implemented

Continuous Bias Monitoring

Ongoing monitoring for AI bias across participant demographics with automated alerting for disparate impact

AI-4
Implemented

AI Explainability and Transparency

Plain-language explanations, confidence scores, and source data transparency for all AI recommendations

AI-5
Implemented

Human-in-the-Loop Design

AI provides insights while healthcare professionals maintain ultimate decision-making authority

AI-6
Implemented

AI Safety Guardrails

Multiple safety mechanisms including graceful degradation, unusual pattern alerts, and manual override capabilities

AI-7
Implemented

No External AI Data Sharing

Participant data never transmitted to external AI services (OpenAI, Anthropic, etc.) - all processing remains within our dedicated secure GCP environment

AI-8
Implemented

AI Performance Monitoring

Continuous monitoring of AI system performance with automated alerts for accuracy degradation or unusual behavior

Security Controls Summary

8
Infrastructure Security
8
Organizational Security
5
Product Security
9
Internal Procedures
3
Data & Privacy
8
AI Governance
41
Total Controls Implemented
100% Implementation Rate

Subprocessors & Data Partners

We maintain strict control over data access and work only with trusted partners who meet our security standards.

Infrastructure Partner

Google Cloud Platform

BAA Signed
Services Provided
  • • Cloud infrastructure hosting
  • • Database services (encrypted)
  • • Network security and DDoS protection
  • • Backup and disaster recovery
Data Access

Google Cloud does not have access to participant data. All data is encrypted with customer-managed keys, and Google personnel cannot access unencrypted data under any circumstances.

Compliance Certifications
HIPAA Compliant
SOC 2 Type II
ISO 27001
FedRAMP

Data Processing Commitment

Private AI Processing

AI models run within our dedicated secure GCP environment

Zero External Access

No participant data shared with external AI services

Full Encryption

End-to-end encryption at all times

Questions About Our AI Ethics or Compliance?

Our compliance team is here to answer your questions about our AI governance, security controls, and ethical commitments.