AI & Automated Decision-Making Transparency Notice

1. Overview

MayDay-IC uses artificial intelligence (AI) and automated decision-making technologies to assist emergency responders in performing their duties more effectively. AI provides recommendations only; all clinical and operational decisions are made by trained personnel. No AI system within MayDay-IC operates autonomously or makes binding decisions without human oversight and approval.

This transparency notice describes the AI systems we use, how they work, what data they process, how we ensure fairness and accuracy, and your rights regarding AI-assisted features.

2. AI Systems in Use

The following table details each AI-powered feature within MayDay-IC, including the data it processes and the role of human oversight:

System Name	Purpose	Data Used	Output Type	Human Override
Triage Recommendation Engine	Suggests triage categories based on patient vitals	Vitals, symptoms, age	Triage category suggestion	Always — responder makes final call
Protocol Advisor	Analyzes agency protocols against national standards	Uploaded PDF protocols	Analysis report	Always
After-Action Report Generator	Drafts post-incident reports	Incident data, timestamps, actions	Draft report text	Always — requires human review
Receipt Scanner	Extracts billing data from receipt photos	Receipt images	Structured cost data	Always — human verifies
PTT Transcription	Transcribes push-to-talk audio	Audio recordings	Text transcript	Always
Evacuation Address Extraction	Parses addresses from evacuation data	Text/document input	Structured addresses	Always
Live AI Translator	Real-time language translation	Text input	Translated text	Always
Training Exercise Analysis	Evaluates crew performance in exercises	Exercise data, actions	Performance analysis	Always

3. AI Provider

MayDay-IC uses OpenAI (GPT models) as our primary AI provider. Our relationship with OpenAI is governed by a Data Processing Agreement (DPA) that ensures:

• No training on our data: OpenAI does not use any data submitted through MayDay-IC to train, improve, or develop their AI models.
• Data processing controls: All data is processed in accordance with our DPA and applicable privacy regulations including HIPAA and GDPR.
• Data retention: OpenAI retains API inputs and outputs for up to 30 days solely for abuse and misuse monitoring, after which they are automatically deleted.
• Security: Data is encrypted in transit (TLS 1.2+) and at rest. OpenAI maintains SOC 2 Type II compliance.

For more information, see the OpenAI Privacy Policy.

4. GDPR Article 22 Compliance

Under Article 22 of the General Data Protection Regulation (GDPR), individuals have the right not to be subject to decisions based solely on automated processing that produce legal effects or similarly significantly affect them.

5. Colorado AI Act Compliance

The Colorado Artificial Intelligence Act requires disclosure of high-risk AI systems. MayDay-IC acknowledges the following:

• High-risk AI identification: The Triage Recommendation Engine may be considered a high-risk AI system under the Colorado AI Act because it provides recommendations related to healthcare triage decisions.
• User notification: Users are clearly notified when they are interacting with an AI-powered feature. AI-generated recommendations are visually distinguished from human-entered data.
• Algorithmic impact assessment: We have performed and maintain an algorithmic impact assessment for our triage recommendation system, evaluating potential risks, biases, and mitigation measures.
• Human oversight: All triage recommendations require confirmation by a trained responder. The AI suggestion does not automatically assign a triage category.
• Opt-out availability: Users may choose to use manual triage workflows instead of AI-assisted recommendations at any time.

6. EU AI Act Considerations

Under the European Union Artificial Intelligence Act, AI systems used in healthcare contexts may be classified as high-risk. MayDay-IC addresses these requirements as follows:

• Risk classification: Our Triage Recommendation Engine is classified as a potential high-risk AI system in the healthcare context. We treat it accordingly with enhanced oversight and documentation.
• Human oversight: All AI-assisted triage decisions require mandatory human review and confirmation by qualified emergency medical personnel.
• Accuracy and robustness: Our triage algorithms are based on nationally recognized standards (START, JumpSTART, SALT, ESI) and are regularly validated against clinical outcomes.
• Transparency: Users are informed when AI is providing recommendations, and the basis for those recommendations is disclosed.
• Documentation: We maintain comprehensive technical documentation of our AI systems, including training data sources, model architecture descriptions, and performance metrics.
• Post-market monitoring: We continuously monitor AI system performance and collect feedback from users to identify and address issues.

7. Data Handling

We take rigorous steps to protect sensitive data when using AI features:

• PHI minimization: AI prompts are constructed to avoid including raw Protected Health Information (PHI) when possible. Data is de-identified or abstracted before being sent to AI providers.
• No model training: Data sent to OpenAI through MayDay-IC is not used for training, improving, or developing AI models.
• Limited retention: AI responses are not stored by OpenAI beyond their 30-day abuse monitoring window. Within MayDay-IC, AI outputs are stored only as part of the incident record for operational and compliance purposes.
• Encryption: All data transmitted to and from AI providers is encrypted using TLS 1.2 or higher.
• Access controls: Only authorized users with appropriate roles can access AI-powered features. All AI interactions are logged in the audit trail.
• HIPAA compliance: Our use of AI is covered under our Business Associate Agreement (BAA) with OpenAI, ensuring HIPAA-compliant data handling.

8. Limitations & Disclaimers

9. Bias & Fairness

We are committed to ensuring our AI systems operate fairly and without discrimination:

• Bias monitoring: We regularly monitor AI outputs for potential biases related to age, gender, race, ethnicity, or other protected characteristics.
• Evidence-based standards: Our triage algorithms follow nationally recognized, evidence-based standards including START (Simple Triage and Rapid Treatment), JumpSTART (pediatric triage), SALT (Sort, Assess, Lifesaving interventions, Treatment/transport), and ESI (Emergency Severity Index).
• No profiling: We do not use AI for profiling individuals or making discriminatory determinations based on personal characteristics.
• Equitable access: AI features are available to all authorized users regardless of their demographic characteristics or the demographics of the patients they serve.
• Feedback mechanisms: Users can report concerns about AI bias or fairness through our feedback channels, and all reports are investigated by our AI ethics team.
• Regular audits: We conduct periodic audits of AI system outputs to identify and remediate any emerging patterns of bias.

10. Your Rights

As a user of MayDay-IC, you have the following rights regarding AI-assisted features:

• Request human review: You may request that any AI-generated recommendation or output be reviewed by a human. All AI outputs are already subject to mandatory human review before action is taken.
• Opt out of AI features: You may choose to use manual workflows instead of AI-assisted features at any time. The core functionality of MayDay-IC is available without AI.
• Access information about AI logic: You may request information about the logic, significance, and envisaged consequences of AI processing that affects you, as provided under GDPR Article 13(2)(f) and Article 15(1)(h).
• Lodge complaints: You may lodge complaints about AI outputs, potential bias, or any concerns through our feedback channels or directly to our AI ethics team.
• Data subject rights: Your existing data protection rights (access, rectification, erasure, portability) apply to data processed by AI systems. See our Privacy Policy for full details.

11. Contact Us