Our Approach to Building Safer AI Experiences for Families
At EBLOMA, safety has always been a core design principle. Our AI companions— including WITPAW and future products —are built with a layered protection strategy aimed at reducing inappropriate or harmful responses.
Recent incidents in the AI toy industry highlight the importance of child safety and content moderation. EBLOMA adheres to rigorous standards to ensure a safe experience for children.We want to explain clearly, and transparently, what EBLOMA is doing today
1. What We Are Doing Today
(1) Application Level Filtering
We use a combination of:
Keyword filtering (including sensitive, violent, adult, or harmful terms)
Topic blocking
Character-limited responses
This ensures that the AI does not freely generate mature or unsafe content.
(2) Controlled Character Personalities
Each EBLOMA character speaks within:
A defined personality
A simple vocabulary
Age-appropriate tone
This prevents unexpected behavior or improvisation outside the intended design.
(3) Intermediary Safety Layer
For devices built through the platform, we configure:
Safety rules
Topic restrictions
Prompt filtering
Response moderation
This adds an additional barrier before content reaches children.
2. Our Philosophy: Child Safety First
We believe AI toys should:
Avoid mature topics
Communicate gently
Encourage imagination
Protect emotional wellbeing
Give parents confidence
While no AI system is infallible, our priority is to minimize risks wherever possible, and to continue building safer layers over time.
3. Continuous Improvement
We are actively:
Testing broader safety filters
Expanding restricted vocabulary lists
Monitoring real-world usage
Refining character scripts
Evaluating additional safety tools
As regulations evolve and formal standards become more defined for AI toys, we aim to align with them step-by-step.
4. A Message to Families
Thank you for trusting EBLOMA.
We are committed to a responsible development path that prioritizes transparency over exaggeration.
If you have concerns, suggestions, or specific questions about our safety measures, you can always contact us at:
care@ebloma.com
