The Eavesdropping Dilemma
An interactive exploration of data privacy in the age of AI assistants. Discover the trade-offs between convenience and control.
The Privacy Problem
This section explores the core of the privacy challenge: the vast amount of data AI assistants collect and the risks this creates. Understand what your assistant knows about you and how that information can be vulnerable.
AI assistants are fueled by data. This data comes from what you say directly, but also from a wide array of passive sources, creating a detailed, and often surprisingly intimate, profile of your life.
🗣️ Voice & Transcripts
Audio of your commands and surrounding conversations from "false wakes," which are often stored indefinitely.
📍 Location Data
Precise GPS data provides context for requests but also tracks your movements for profiling.
👤 Biometric Data
Your unique voiceprint for speaker identification, which can also be used for other authentication purposes.
💻 Device & Usage Data
IP addresses, device serial numbers, and logs of your interaction patterns and feature usage.
🌐 Web & Inferred Data
Browse habits and, critically, sensitive inferences about your health, politics, and income, even if never stated.
⚙️ Sensor Data
Information from accelerometers, gyroscopes, and light sensors can infer activities like walking, driving, or sleeping.
The Players: A Company Comparison
The major voice assistants operate under different business models, which directly influences their approach to data collection. This chart compares the number of data points collected, while the cards below reveal specific company practices and controversies.
Amazon Alexa
Commerce-Driven
Google Assistant
Advertising-Centric
Apple Siri
Privacy-as-Feature
The Rules: Global Regulations
This section provides a high-level comparison of the two most influential data privacy laws: Europe's GDPR and California's CCPA/CPRA. See how they approach key issues like automated decision-making and the right to delete your data, and understand the compliance challenges AI companies face.
🇪🇺 GDPR / EU AI Act
🇺🇸 CCPA / CPRA (California)
A key challenge for both frameworks is the "un-baking the cake" problem: legal rights like data deletion are technically infeasible to apply to a fully trained AI model without starting over.
The Solutions: A Path Forward
Addressing AI privacy requires a multi-faceted approach. This section explores cutting-edge Privacy-Enhancing Technologies (PETs) that build protection into system design, followed by actionable recommendations for key stakeholders.