Google confirms that its privacy policy has consistently allowed the use of publicly available web data to train AI models, including the newly launched Bard chatbot. This disclosure clarifies that while private user account data remains protected and separate, the company relies on open internet sources to develop and improve its artificial intelligence technologies. This approach highlights a critical intersection between commercial AI development and open data ecosystems. By leveraging publicly accessible information, Google demonstrates how large-scale language models depend on the vast resources of the open web, raising questions about data scraping practices and the sustainability of relying on unstructured public sources. The article is relevant to open data because it illustrates the real-world implications of open web data for AI training. It underscores the tension between corporate AI advancements and public data governance, emphasizing the need for transparency regarding how open information is utilized and protected in automated systems.
Source: timesofindia.indiatimes.comPublished on 2023-07-07