摘要 |
FIELD: information technologies. ^ SUBSTANCE: agent tracks interactive conversations and may interfere in case the conversation content satisfies specific definable criteria, which demonstrate that the content could have been inappropriate or possibly harmful or dangerous. The agent uses a procedure of natural language processing to assess the conversation content during its entire length and/or conversations for all separate users. Using procedures of natural language processing, the agent may then provide notices or warnings to separate users, when a specific conversation or its part is considered problematic. ^ EFFECT: improved safety of online conversations. ^ 12 cl, 7 dwg |