In Short:
OpenAI advises ChatGPT users to adjust settings to control data usage. Dall-E 3 also allows image removal requests. Perplexity, Quora, Rev, and Slack also use AI for data usage. Opt-out options available on some platforms. Slack uses customer interactions to improve machine learning capabilities. Adobe users on enterprise accounts may have limited options for opting out.
OpenAI Data Controls
OpenAI has provided instructions on how ChatGPT web users without accounts can navigate to Settings and uncheck the Improve the model for everyone option. For users with accounts, the process involves selecting ChatGPT, Settings, Data Controls, and turning off Chat History & Training. Mobile app users need to go to Settings, choose Data Controls, and disable Chat History & Training. It is important to note that changing these settings does not sync across different browsers or devices, so adjustments need to be made everywhere ChatGPT is used.
OpenAI Dall-E 3 Image Generator
OpenAI has introduced a form for users to request the removal of their images from future training datasets. The form requires information such as name, email, image rights ownership, image details, and uploads of the image(s). Additionally, for users with a high volume of online images to be removed from training data, adding GPTBot to the robots.txt file of the hosting website may be a more efficient solution.
Perplexity
Perplexity is a startup utilizing AI to assist in web searches and answering questions. Users are automatically included in the data used to train Perplexity’s AI, but this can be disabled by navigating to the account name, accessing the Account section, and turning off the AI Data Retention toggle.
Quora
Quora mentions that it does not currently employ user answers for training AI, nor has it sold any user data for such purposes. However, users have the option to opt-out of allowing large language models to be trained on their content by visiting the Settings page, selecting Privacy, and disabling the “Allow large language models to be trained on your content” option. While Quora provides opt-out choices, some posts may still be used for training language models if replied to machine-generated answers.
Rev
Rev, a voice transcription service, states that it uses data anonymously and perpetually for training AI systems, even after account deletion. Users with concerns about their data being used can opt-out by sending an email to [email protected], as mentioned in the help pages.
Slack AI
Slack incorporates machine learning in its product features, improving machine-learning capabilities using customer interactions. While Slack does not use customer data to train large language models, interactions are used to enhance software capabilities. For users with enterprise accounts, individual opt-out options are limited based on Slack’s data analysis practices.
