Responsible AI usage
POSTPALS uses a limited amount of artificial intelligence (AI) to help develop our learning experiences while protecting the rights, safety, and privacy of all users. This policy explains how we use AI, the principles guiding its use, and the safeguards we have in place.
How we use AI
We use AI in the following ways, always alongside people.
Science and data unpacking
Working with our institutional Partners, we receive a great deal of data about their mission, their animals, and their science. We use AI and humans to unpack the information, organize it, and plan our Adventure Quests.
Adventure Quest packaging
Once the information has been unpacked and organized, we develop our Explorer’s curriculum and learning themes. We use this to build the learning models and have AI, with human guidance, help to develop Adventure Quests that are informative, enjoyable and fun.
Personalized postcard creation
Our Explorers are assigned a monthly Pal to partner with them in their Adventure Quests. Their Pal may be a shark, a bear, a bird or even a planet. Each month’s Quest is introduced by their Pal, with a postcard. For Parents who choose, we provide the option to personalize the postcard pictures with their Pal and Explorer together. This creative work is done using humans and AI. If preferred, the postcard picture can be with just the Pal.
Principles of responsible AI use
We follow these core principles.
Human oversight is paramount
Humans, both our team and our Partner’s scientists, remain the final decision-makers in assessing the Adventure Quest curriculum and content. AI-generated insights are advisory.
Curated by our team
POSTPALS Adventure Quests are curated by the POSTPALS team (with the participation of our Partner’s scientists). Our content creation team carefully curates and reviews these learning interactions per the modeled Adventure Quest activity to ensure accurate focus on the Explorer’s activities. Defining clear goals allows our models to avoid any off-topic conversation, instead of the other way trying to flag everything considered “bad”.
Privacy & data protection
We collect only the data necessary for our adventure and educational purposes. Personal information is handled in compliance with applicable data protection laws (e.g., COPPA, FERPA, GDPR). No child data is ever used to train, fine-tune, or otherwise improve any external or internal large language models (LLMs).
Security
All AI data processing occurs through secure, encrypted channels.
Bias awareness
We monitor AI outputs for potential bias or inaccuracies and continually improve our models and processes.
Age appropriateness
AI interactions are designed to be safe and suitable for the intended Explorer age group.
What we don’t do
Some lines we will never cross.
No automated decisions
We do not use AI to make automated decisions that directly affect an Explorer’s Adventure Quest without human review. Curriculum and learnings generated should always be considered in the context of AI and never be taken as truth.
No selling data
We do not sell or share Explorer data with advertisers or unrelated third parties.
No profiling
We do not use AI to profile Parents or Explorers for non-learning purposes.
Accountability & updates
We hold ourselves to these standards, and we’ll keep you informed as they evolve.
Accountability & feedback
We regularly audit AI systems for quality, accuracy, and safety. Explorers, Parents, and guardians are encouraged to report any concerns or unexpected content through our Contact page.
Updates to this policy
We may update this policy as our AI technology or regulations evolve. We will notify users of significant changes in advance.
