Generative AI Browsing Obfuscation for Human-Like Privacy Traffic

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for maintaining user privacy on the web, such as HTTPS, VPNs, and browser incognito modes, are insufficient in preventing tracking and profiling by trackers, as they can be circumvented by techniques like fingerprinting, and random noise traffic can be distinguished from real traffic.

Innovation Solution

Utilizing a large or small language model to analyze web content and generate fake, human-like traffic patterns that are negatively correlated with a user's actual behavior, obfuscating specific identity features and categories to prevent profile generation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If random noise traffic is injected into user browsing histories, then user privacy is protected by adding chaos to web traffic behavior, but the random traffic can be distinguished from real traffic through analysis of topic similarity, visit frequency, or link graph traversal

Engineering Contradiction:
Improveprivacy protection effectivenessVSAvoidtraffic authenticity verification
Core Design Contradiction:
ReliabilityVSDifficulty of detecting and measuring

Solution Approach 1:

The system generates synthetic browsing trajectories that copy the statistical patterns and behavioral characteristics of real user navigation. By creating artificial browsing paths that mimic human-like exploration patterns, the system produces noise traffic indistinguishable from genuine user behavior, thereby protecting privacy while maintaining traffic authenticity.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system dynamically adjusts parameters such as visit frequency distributions, topic similarity metrics, and link traversal patterns to match real user behavior statistics. By changing these parameters to align with empirical observations of genuine browsing, the noise traffic becomes statistically indistinguishable from real traffic, resolving the contradiction between privacy protection and detectability.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If trackers collect detailed user engagement data to generate comprehensive user profiles, then commercial advertising value is maximized, but user personal privacy is lost

Engineering Contradiction:
Improveadvertising effectivenessVSAvoiduser privacy
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The system introduces synthetic browsing trajectories as an intermediary layer between the user and trackers. This intermediary generates artificial engagement data that masks the user's true behavior patterns while still providing sufficient signal for advertisers to deliver relevant ads, thereby protecting privacy without completely eliminating advertising effectiveness.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system creates copies of browsing behavior that preserve the statistical properties needed for advertising targeting while obscuring individual user identity. By copying aggregate behavioral patterns rather than individual trajectories, the system maintains advertising productivity while preventing personal profile generation.

Inventive Principle:
Principle #26Copying

3Reliability

If existing privacy methods like HTTPS, VPNs, or incognito mode are used, then some level of encryption or tracking artifact storage is prevented, but tracking is still possible through techniques such as fingerprinting

Engineering Contradiction:
Improveprivacy protection baselineVSAvoidanti-tracking capability
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The system dynamically generates and injects synthetic browsing trajectories that adapt to the user's actual browsing behavior in real-time. This dynamic approach creates ever-changing noise patterns that are difficult for trackers to distinguish from genuine user activity, providing superior adaptability compared to static privacy methods while maintaining strong baseline protection.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4672062A1Generative ai based user behavior obfuscation
Publication Date: 2025.12.31 INTEL CORP
  • EP4672062A1 patent drawingFigure 1~2
  • EP4672062A1 patent drawingFigure 3
  • EP4672062A1 patent drawingFigure 4~5

AI summary

A device includes a memory, including a first artificial neural network; and a second artificial neural network; and a processor, configured to implement the first artificial neural network to generate a first output based on a first input; implement the second artificial neural network to generate second output based on a second input; wherein the first input includes or represents content of a website; wherein the first output includes a content category associated with the content of the website; wherein the second input includes the content category; and wherein the second output is negatively correlated with the content category of the second input.