Facebook will soon become the newest platform to train its AI models using user-generated posts and images.
Opting out is a possibility, but it seems that Meta purposefully made the procedure challenging and time-consuming. It’s also prone to failure.
On June 26, 2024, Facebook’s privacy policies will shift, using public user content instead of private messages to train and enhance its AI services. Although there is a method to object to the data scraping through Meta, it is more involved than just checking a box.
Go to this help page and select the option labeled “I want to object to or restrict the processing of my personal information from third parties used for building and improving AI at Meta” if you would like to opt-out.
It’s then a matter of filling in your country of residence, name, surname, and email address. The form inquires as to whether your information is known to Meta’s AI models. This means responding to any inquiries you made that resulted in the exposure of your personal information in an AI response at the meta model, feature, or experience. Meta also asks for a screenshot proving your personal information appeared in a response from one of its AI models.
The last section is an additional context box that asks for a brief explanation of your concern and what you’re requesting.
X user Tantiiacrul writes that Meta is also sending out notifications to users about the new AI features. You can click on this to be directed to a similar form with a section on “how this processing impacts you,” after which there is a “right to object” link. Meta also requires users to enter an OTP sent to their email address. The business says this is necessary for security, but the fact that the submission will not be made until the code is entered raises suspicions that this is just another ruse to discourage users from completing the form.
2. I should start by mentioning that I’ve worked in growth teams who conduct experiments to minimise friction for over a decade and I know how to streamline an experience. Rule: every additional step you add dramatically decrease the % of people who’ll make it through to the end.
– Tantacrul (@Tantacrul) May 26, 2024
“If your objection is upheld, it will be implemented moving forward,” reads Meta’s notice.
Although many of his friends and Reddit users who also objected received an error message when attempting to submit the form, Tantacrul writes that the company did honor his objection.
These days, using user content scraping to train AI models is becoming commonplace. In order to integrate user posts into ChatGPT, OpenAI partnered with Reddit earlier this month.
