Modern vehicles offer drivers and passengers a range of exciting innovations – from giant touchscreens to near-independent driving capabilities. But one increasingly common innovation raises significant questions about the right priorities between safety and privacy. We used ChatGPT to explain the struggle.
Many cars today have a huge number of sensors of various types, including cameras, in order to improve the driving experience by leveraging information from the vehicle's environment. Some of these cameras are pointed outward, to help reverse or warn of thieves and car shorts – but some are pointed inward, follow the driver's gaze, and know how to cross-reference the information they collect with the vehicle's speed, driving nature, etc.
These technologies serve a multilateral purpose – from starting the car only when the owner's face is recognized in the driver's seat, to identifying distractions and "pecking" while driving. This data is often collected, stored, and also transferred to the manufacturer for various reasons, such as improving driving experiences or offering personalized services.
Alongside the benefits, the proliferation of computer and tracking systems in the vehicle is a potential nightmare for privacy. The data collected – including driver identity, driving habits and places you visit most often – reveals a detailed profile. A report by the Mozilla Foundation (the one behind the Firefox browser) called vehicles "the worst product category for privacy" due to concerns about lax privacy policies among automakers.
To address privacy concerns related to driver data, researchers are exploring innovative approaches. Distributed machine learning is considered a promising idea, allowing algorithms to learn from data on local devices without transferring the data to a central server. This approach, implemented in Google's keyboard app, allows it to offer custom autocompletes without compromising user privacy. A similar but slightly different way proposed is blockchain-based distributed machine learning, in which there is no identification of the users from whom the data was collected. Techniques like location obfuscation or altering location data contribute to the arsenal of privacy measures.
Until the field begins to implement such methods, drivers with modern cars need to develop awareness and think twice before purchasing cars with advanced features based on what might be called espionage, or before employing those options.
Wrong? We'll fix it! If you find a mistake in the article, please share with us