‘We believe this is a practical step toward an open web’: Don't want AI companies training on your website? OpenAI, Google and Microsoft are backing a new way to opt out

https://cdn.mos.cms.futurecdn.net/8iN8JcsjaeEppccYgURuY9-1920-80.jpg

OpenAI, Google, Microsoft and Anthropic have agreed to support a new way for website owners to stop their content being used to train AI models. The companies are backing a shared opt-out signal in a website’s robots.txt file, which tells AI crawlers whether they're allowed to use its content for training.

Like a modern day Library of Alexandria, the internet is the single largest source of human knowledge that exists. Countless people have spent their lives cataloguing and uploading the entire history of humanity for the betterment of everyone.

AI companies are a tinderbox in this regard. To train and develop better capabilities for their models, AI companies are repeatedly scraping every corner of the internet to create archives of training data. Compounded by changes in search and how people seek information, AI threatens to become the single point of access to information.

But just because you have...

Copyright of this story solely belongs to www.techradar.com. To see the full text click HERE

Read more