Google Now Allowing Website Owners to Opt Out of Google Bard

Google has introduced a mechanism for web publishers to selectively opt out of their generative AI crawler, Google-Extended, which powers services like Bard. This update addresses concerns regarding data privacy and intellectual property by allowing site owners to prevent their content from being used to train AI models without disabling the standard search crawler. This distinction is crucial for maintaining search engine visibility while protecting proprietary information, a significant shift from previous options that forced publishers to block all access. The relevance to open data lies in the tension between training large language models on publicly available information and the emerging right to opt out. As AI companies increasingly rely on web scraping to fuel their systems, this development highlights the growing need for granular control over data usage. It forces a re-evaluation of what constitutes "public" data when the owners of that data explicitly wish to restrict its use for commercial AI training, challenging the default assumption of open accessibility. Furthermore, this tool ensures that publishers can halt not just current AI products like Vertex AI but also future generations, providing long-term safeguards. This evolution in web publisher controls signifies a broader industry trend where the ease of data access is being balanced against the ethical and legal demands of content creators. It underscores that open data practices are evolving to include explicit consent mechanisms, fundamentally altering how AI developers can interact with public web content.

Source: techtimes.com
Published on 2023-09-30

Google Lets Publishers Opt Out of AI Training Data »

Open Data

Related news