Apple Intelligence is now accessible to users around the world, with a broader rollout planned for later this year. The AI toolkit is also expected to be the selling point for the upcoming iPhone 16 series. However, Apple"s efforts to make a big splash with Apple Intelligence are already hobbled by major websites and social platforms that are blocking Apple"s AI bot from scraping data.
As reported by Wired, a while ago, Apple revealed a tool for publishers to let them opt out of the company"s AI training program. This means Apple"s AI bot can no longer scrap data for training purposes. The impact is profound, with some of the biggest news websites and social platforms opting out of the Apple AI program, thereby impacting the landscape of Apple AI development.
According to the report, Facebook, Instagram, Craigslist, Tumblr, The New York Times, The Financial Times, The Atlantic, Vox Media, the USA Today network, and WIRED"s parent company, Condé Nast are among the websites that have seemingly opted out of Apple Intelligence training.
AI chatbots like ChatGPT rely heavily on the valuable content published on the Internet to enrich their knowledge base and provide users with correct answers. However, publishers, as the backbone of this model, are not often happy with this arrangement as AI chatbots consume the content for free. While chatbot developers have tried to mitigate the problem by adding source links to the end of each prompt, publishers still rightfully seek compensation for their content.
Since 2015, Apple has been using a tool called "Applebot" to crawl the Internet and feed products such as Siri and Spotlight. The Applebot, however, is now used to train Apple Intelligence. By publishers blocking the Applebot, their content won"t be used to train Apple"s large language models, but the content will still be used for Siri and Spotlight searches.
After all, money is probably the main reason these publishers have opted out of Apple Intelligence training. If Apple truly respects the publishers" choice, it needs to either sign a business agreement with them to consume their data or find other sources to feed its large language models.