Google is accused of “stealing everything ever shared on the internet,” to fuel its AI, including copyrighted works and personal data. The case comes after a privacy policy update spotted by Gizmodo which says the company scrapes the web’s public information.

A 6-year-old, a best-selling author, and others accuse Google of stealing “everything ever shared on the internet" after Gizmodo noted a privacy policy change.

  • @RivalarrivalEnglish
    8•1 year ago

    Reading or other forms of consumption sure, but copying — no.

    Scraping for purposes of indexing a search engine actually requires copying, and presenting those search results actually requires re-distributing portions of the copyrighted work. Search engines have been using various forms of “artificial intelligence” for decades, and their models have survived countless legal challenges.

    Training an LLM would be considered an act of consumption, not an act of copying. It is less infringing than indexing, in that the LLM is designed to not regurgitate copies of existing work, but rather to produce novel content.

    When I post something online, say a trip report on my personal travel blog, does that mean I consent to ie Google using my intellectual property for training their LLM? My answer is a resounding no.

    “Your answer” is wrong.

    Google is a member of the public. Microsoft is a member of the public. Meta is a member of the public. Your pervy neighbor is a member of the public. I am a member of the public, as are all your social media friends. If you want to prohibit a member of the public from consuming your content, you cannot post it in a publicly-accessible forum.