Interpreting Privacy Policies with Artificial Intelligence

A team of researchers from EPFL in Switzerland, University of Wisconsin and University of Michigan developed a deep learning-based program that can…

NVIDIA logo
NVIDIA
•
Brad Nemire
•2 min read•intermediate•
--
•View Original
Artificial IntelligenceDeep LearningTensorFlow

Overview

A team of researchers from EPFL, University of Wisconsin, and University of Michigan developed a deep learning-based program named Polisis that can automatically interpret online privacy policies. The program utilizes graphs and color codes to present data usage information clearly, trained on over 130,000 privacy policies from the Google Play Store.

What You'll Learn

1

How to use Polisis to analyze privacy policies of online services

2

Why understanding privacy policies is crucial for data protection

3

When to utilize AI tools for interpreting complex legal documents

Key Questions Answered

How does Polisis help users understand privacy policies?
Polisis uses deep learning to analyze privacy policies and presents the findings through simple graphs and color codes, making it easier for users to see how their data might be used. This tool can quickly interpret complex legal language and highlight important clauses related to data protection.
What technologies were used to develop Polisis?
The researchers utilized TITAN X GPUs and the cuDNN-accelerated TensorFlow deep learning framework to train their convolutional neural networks on over 130,000 online privacy policies. This technological foundation enabled the rapid analysis and visualization of privacy information.
What is the purpose of the PriBot chatbot?
The PriBot chatbot allows users to ask specific questions about a website's data protection policy, providing an interactive way to understand privacy practices. This tool complements the Polisis program by offering personalized insights into privacy concerns.

Key Statistics & Figures

Number of online privacy policies analyzed
130,000
This extensive dataset was used to train the convolutional neural networks for the Polisis program.

Technologies & Tools

Some links below are affiliate links. We may earn a commission if you make a purchase.

Hardware
Titan X Gpus
Used for training the deep learning models.
Software
Cudnn
Accelerates the TensorFlow deep learning framework.
Software
Tensorflow
Framework used for developing the deep learning models.

Key Actionable Insights

1
Utilize the Polisis browser extension to quickly analyze privacy policies before using online services.
This tool can help users make informed decisions about their data privacy by clearly displaying how their information may be used, especially for services that are not transparent about their data practices.
2
Engage with the PriBot chatbot for personalized inquiries regarding privacy policies.
This interactive feature allows users to clarify specific concerns they may have about data protection, enhancing their understanding of the implications of using various online services.
3
Leverage deep learning frameworks like TensorFlow for similar projects in data interpretation.
By adopting advanced technologies, developers can create tools that simplify complex information, making it accessible to a broader audience, similar to what Polisis achieves with privacy policies.