Meta has announced plans to make Europeans – specifically their data – work harder for Meta’s AI models.
But the giant is presenting these plans as making its AI “work harder for Europeans.”
Meta’s AI models will soon start to be trained with data from EU users over the age of 18, including their public posts, comments, and chat history – such as questions and queries – with Meta AI. However, Meta will keep its hands-off private messages “with friends and family.”
Meta AI was rolled out in Europe in March, and the company is moving fast to make use of existing and new data now at its disposal. A blog post presented the purpose of Meta AI as the desire to “better support millions of people and businesses in the EU.”
Meanwhile, the need to hoover up data belonging to those people and businesses into Meta AI is allegedly necessary “to better understand and reflect their cultures, languages, and history,” not to mention, “improve user experience.”
Meta is either extremely ambitious and optimistic regarding its models, or eager to give off that impression, when it claims the AI technology will be sophisticated enough to differentiate between dialects and even colloquialisms, and the way humor and sarcasm are expressed across EU’s dozens of countries.
And when the scheme launches (which is yet to happen) everybody will be opted-in by default, and will have to fill out an “objection form” found on the privacy policy pages in order to opt-out. Starting this week, users will be receiving emails and in-app notifications informing them about the new scheme, which will include the link to the opt-out form.
Meta already tried to use EU and British data to train AI last year, but both jurisdictions have stricter data protection laws than the US and this resulted in delays.
This latest attempt appears to still be facing some hurdles from regulators, although the company on Monday seemed confident the project will be able to proceed “soon.”
Meta made a note of the fact it put user data to work in other regions immediately after Meta AI was launched, as well as that they are now following in the footsteps of Google and OpenAI – “both of which have already used data from European users to train their AI models.”
