Google Data-Driven Approach: Training Bard and AI Models with User Data

AI-powered chatbots like Google Bard and OpenAI ChatGPT rely on data to improve their performance.

Twitter imposed read limits to prevent AI companies from scraping data for training their models. Google has since made it clear that in order to train Bard and other generative AI products, it will make use of publicly accessible data.

In a recent privacy update, Google stated that it uses information that is already in the public domain to improve its services and create new ones like Google Translate, Bard, and Cloud AI capabilities.

It emphasizes that it does not use user’s private data, but rather relies on data that is accessible to the public.

According to Google statement, the purpose of collecting data is to provide better services to its users. This involves understanding users’ language preferences, identifying useful advertisements, and determining the online connections that matter to them.

Google’s privacy policy emphasizes that the type of data collected and its usage depend on how users engage with the services and manage their privacy controls.

The collected data will include your device type, browser type, operating system, mobile network information (such as carrier name and phone number), and the version number of your application.

It also encompasses the interaction between users’ apps, browsers, and devices with Google’s services, including IP addresses, crash reports, system activity, and request details.

In conclusion, Google will use publicly available data but not personal user data to train its AI models and improve its services. The collected data includes various details about devices, operating systems, and user interactions with Google’s services.