At Chronus, we are committed to the responsible and transparent use of artificial intelligence. This page outlines how AI is used in our AI features, what data is processed, how we protect user privacy, and what controls are available. Your mentoring program administrators determines which AI features are enabled for the particular program you are enrolled.
Feature Overview
Chronus AI-powered features are designed to help users clarify goals, receive personalized guidance, and enhance engagement throughout their learning journey. Depending on the feature, AI may generate:
Short narrative summaries or insights.
Actionable focus areas or suggestions.
Personalized recommendations based on user input.
Summaries and transcripts of meetings between mentors/mentees
Suggested next actions
Mentorship goals
Interactive roleplay practice with an AI persona, along with a structured debrief
These features are meant to complement human judgment, not replace it, ensuring that people remain in control of their learning journey.
AI Usage Disclosure
Chronus uses OpenAI’s Whisper for speech to text purposes and GPT‑4o & 5.2 language model via API to power other AI features. The AI processes meeting audio, meeting metadata (participants, date/time), meeting history, program context, participant goals/tasks, previous meeting transcripts, user profile fields (as determined by program administrator), user focus areas / game plan / success statement / goals / unused recommendations, user-provided text input to generate relevant and structured content in real time.
AI Roleplays use Anthropic's Claude Sonnet model to power persona responses and session feedback, and ElevenLabs for speech to text and text to speech in voice sessions. The AI processes the roleplay scenario, persona details, and user input (text or voice) to generate responses and feedback in real time.
All AI-generated content is clearly labeled within the platform. Users are notified when:
An AI Notetaker is included in a meeting.
When a RolePlay is initiated and session feedback is provided
Entering information into AI-enabled fields.
A recommendation is presented to an user or an admin
Additionally, in admin configuration settings where AI features are managed.
Data Handling & Privacy
Data Shared with AI Providers:
User-submitted responses (e.g., about goals, focus areas).
Program metadata: language, mentoring program name.
Mentoring meeting details, including meeting audio, meeting metadata (participants, date/time), meeting history, and previous meeting transcripts.
User goals/tasks, user profile fields (as determined by program administrator), user focus areas / game plan / success statement / goals / unused recommendations, user-provided text input to generate relevant and structured content in real time.
AI Roleplay scenario details and user input (text or voice) during roleplay sessions.
Data Not Shared:
No sensitive personally identifiable information is intentionally shared (e.g., name, email, demographics).
No analytics data, sentiment scores, or admin‑generated metadata.
Important: Chronus applies automated PII redaction to text-based AI features before content is sent to the AI provider. Voice interactions, including speech to text and text to speech, are processed in real time as spoken. Users are required to limit the information they share and informed not to enter sensitive personally identifiable information, in text or voice, to what is required by the application or feature.
Security Measures
In addition to our technical, organizational, and administrative safeguards, AI Providers does not use any inputs received via Chronus to train their models.
Model Behavior & Learning
The underlying AI model are static and remain unchanged by past inputs.
Chronus may update prompts (instructions sent to the model) to improve output quality, but these changes are controlled by Chronus, not the AI.
Oversight & Quality Control
AI-generated content is delivered instantly and is editable by the user, where applicable.
Prompt design includes guardrails to ensure outputs are clear, professional, and appropriate.
Limitations & Accuracy
AI outputs are designed to assist users; however, they may not always be fully accurate or personalised. Users are encouraged to review and adjust AI-generated content to better reflect their individual goals or needs.
Questions?
If you have questions about how AI is used at Chronus, please contact us at [email protected]