Selective Text Prediction for Sensitive Messaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic messaging systems face inefficiencies in communication due to cumbersome drafting and editing processes, which can lead to reduced use of electronic messages, and there is a need to avoid suggesting potentially offensive text predictions, especially in sensitive topics.
Innovation Solution
A computing system that predicts text suggestions based on a corpus of text using machine learning models but refrains from outputting suggestions when an electronic message is determined to be related to sensitive topics, ensuring user control over information analysis and privacy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If text prediction is provided continuously to assist message drafting, then user interaction efficiency is improved, but risk of suggesting offensive or inappropriate text increases
Solution Approach 1:
The system dynamically adjusts the text prediction function based on message sensitivity. When a message is detected as sensitive, the prediction function is temporarily disabled or modified to avoid suggesting inappropriate completions. This dynamic switching allows the system to maintain high productivity for normal messages while preventing harmful suggestions in sensitive contexts.
Solution Approach 2:
The system implements feedback loops where user responses to predictions are monitored and analyzed. When users consistently reject certain types of predictions or when sensitive topics are detected in the message content, the system adjusts its prediction behavior accordingly. This feedback mechanism helps the system learn from user preferences and avoid generating offensive suggestions while maintaining helpful predictions for appropriate contexts.
2Loss of time
If text prediction suggestions are always provided, then drafting speed is improved, but processor resource consumption increases
Solution Approach 1:
Instead of continuously generating predictions for every keystroke, the system employs periodic action by only activating text prediction when certain conditions are met (e.g., after specific keywords, when confidence thresholds are met, or when message sensitivity is low). This periodic activation significantly reduces processor cycles while maintaining drafting speed benefits when predictions are actually provided.
Solution Approach 2:
The system changes operational parameters based on message context. When a message is determined to be sensitive, the prediction parameter is adjusted to disable or reduce prediction frequency. For non-sensitive messages, prediction parameters are optimized for speed. This parameter adaptation allows the system to minimize processor consumption while maintaining high drafting speed when appropriate.
Data Source
AI summary
A computing system is described that includes user interface components configured to receive typed user input; and one or more processors. The one or more processors are configured to: receive, by a computing system and at a first time, a first portion of text typed by a user in an electronic message being edited; predict, based on the first portion of text, a first candidate portion of text to follow the first portion of text; output, for display, the predicted first candidate portion of text for optional selection to append to the first portion of text; determine, at a second time that is after the first time, that the electronic message is directed to a sensitive topic; and responsive to determining that the electronic message is directed to a sensitive topic, refrain from outputting subsequent candidate portions of text for optional selection to append to text in the electronic message.