Tech companies are increasingly turning to historical libraries for AI training data, with Harvard releasing a collection of nearly 1 million public domain books spanning centuries, as well as old newspapers and documents. This initiative, supported by donations from Microsoft and OpenAI, aims to improve AI models and address concerns over copyright infringement by using legally accessible materials while also recognizing the potential risks of outdated or harmful content in the data.
This is an ainewsarticles.com news flash; the original news article can be found here: Read the Full Article…
