Google confirms scraping public data for training Bard

Google has admitted to data scraping from public sources to train its artificial intelligence and services, as going by a recent update to its privacy policy. Spotted first by The Register, Google, in its privacy policy under the research and development column, has mentioned that publicly available information is used to train products and features like Google Translate, Bard, and other Cloud AI capabilities. “Google uses information to improve our services and to develop new products, features and technologies that benefit our users and the public.”

A Google spokesperson said that the company has been transparent about the use of publicly available information to train its models. The new update merely includes Bard and other new services to the list of models being trained this way.

Recently, Twitter owner Elon Musk said that the microblogging platform will place a daily tweet reading limit. He said that this was being done to “address extreme levels of data scraping & system manipulation.”

Another social media platform Reddit has imposed a fee for third-party application program interface (API) access. In an earlier interview with The New York Times, Steve Huffman, Reddit founder and chief executive officer said that the platform’s data corpus is really valuable to be given for free to some of the largest companies. Similarly, companies like Twitter and Stack Overflow have introduced new rules for their API access. 

ChatGPT-maker OpenAI found itself embroiled in fresh legal trouble when a California-based law firm launched a class-action lawsuit against it. As per June 28 The Washington Post report, the law firm has alleged that the AI company has violated copyright rules and the privacy of the public by using data scraped from the internet to train its models.