Social platform X (formerly Twitter) began using their posts and conversations with the AI chatbot Grok to train the latter without notifying users, writes 9to5Mac. In settings, permission to use training data is enabled by default, but users can disable this action, but only in the web version of X.
Image source: 9to5Mac
On Friday, X users noticed that the platform had added a new “Data Sharing” toggle in Settings, explaining that any communications made on the platform, as well as interactions, inputs, and outputs using Grok, are now used for “training and fine-tuning” “
Up until this point, Grok had only been trained on a corpus of publicly available text from the internet, with capabilities and datasets that were a far cry from ChatGPT or Copilot. Social Network X says Grok’s real-time access to X’s public messages allows the chatbot to respond to user queries with relevant information and ideas on a wider range of topics.
According to the developers of the social network, training will allow:
- Improve Grok’s understanding of human language and communication;
- Improve Grok’s ability to provide accurate, relevant and engaging answers;
- Will develop Grok’s sense of humor and wit to make interacting with him more enjoyable;
- Ensure that Grok remains politically impartial and provides balanced responses.
To opt out of using your data for Grok training, you must select the “Privacy and Security” section in the settings of the web version of the X application. Then, under the “Sharing and Personalization” subsection, you need to select Grok and disable the option “Allow your messages, as well as your interactions, inputs, and results, to be used with Grok for training and fine-tuning.”