Talkie
Customer Type
AI Companionship
Talkie is a leading AI-native content community and social platform.
Customer Introduction
Talkie is a leading AI-native content community and social platform where users can engage in real-time dialogues, voice calls, and role-playing with thousands of AI-driven virtual characters (NPCs). The platform's core experience is built upon highly autonomous AIGC interactions, including AI text responses, intelligent voice calls, and the creation of personalized NPC cards via user-generated prompts.
Scenario
In its global operations, Talkie faces unique regulatory and operational challenges inherent to AIGC:
1. AIGC Dialogue Risks: AI-generated text (aimessage) or voice (aiphone) outputs may produce compliance risks such as sensitive political content, terrorism, pornography, or abusive language.
2. NPC Creation Compliance: NPC images generated via user "image_prompts" may contain NSFW content, nudity, child safety violations, or violent imagery, requiring strict alignment with platform guidelines and regulatory baselines.
3. Global Multilingual Pressure: Operating across English, Spanish, Portuguese, Japanese, and Chinese, the platform must balance "freedom of speech" in overseas markets with rigorous interception of P0 high-risk content (e.g., racial discrimination, child protection).
Solutions
Deepcleer provides Talkie with a closed-loop content safety solution integrating AI and human moderation:
1. Comprehensive AIGC Dialogue Defense: Full-volume, real-time detection for AI-generated text and voice calls. Deep recognition is used for English, while a "Translation + Policy Audit" model is applied to Spanish, Portuguese, and Japanese for global consistency.
2. NPC Visual Safety Check: Detection is integrated directly into the image generation phase; if the AI audit issues a "REJECT," the generation process is terminated immediately. Passed images proceed to a human audit for style judgment and operational tagging, balancing compliance with aesthetics.
3. Differentiated Regulatory Strategy: Tailored for global markets, the strategy focuses on intercepting risks at the model output stage while maintaining flexibility for user inputs to balance compliance with user experience.
4. High-Efficiency Quality Assurance: Established a routine inspection and evaluation mechanism, performing large-scale daily spot checks on facial data and weekly evaluations of false positives to drive continuous policy iteration.
Effect & Value
4% → 0.2%
False Negative Optimization
20% → 7%
