In an ongoing regulatory battle with European privacy authorities Announce On Thursday, its AI-powered virtual assistant, Meta AI, was finally launched in the EU. The chatbot-like tool will be launched in Meta’s social platform portfolio, although it has more limited feature settings than it does in the domestic U.S. market.
Additionally, Meta has confirmed to TechCrunch that Meta AI will arrive at WhatsApp in the UK so far since it launched in October, and so far it is limited to Facebook, Instagram and Ray-Ban Meta glasses.
Meta AI already exists in the United States Since 2023acting as an AI assistant, not only chatting and answering questions, but also generating images and creating Style selfie,other Creative miracle. These features are not available in the European version.
Last month, the chat-based version also Landing in a few countries throughout the Middle East and Africa. And now, starting this week, Meta AI will launch all 27 EU countries, plus 14 European countries (and 21 overseas territories), including Iceland, Norway, Serbia and Switzerland.
In addition to being able to chat one-on-one with assistants in various Meta’s applications, Meta AI will also be available in group chats, although the launch conferences will be interleaved – the feature plan first appeared on WhatsApp, including in the EU and the UK – already noted –
Privacy back
The launch of Meta AI by the EU marks the latest effort by Facebook’s parent company to spread its AI across the group as it faces regulatory issues that attack user data training AI models.
And Yuan has Training its AI For years, the technology giant ( Others) Due to the comprehensiveness of the group Privacy Regulations – Including General Data Protection Regulation (GDPR) – This means it needs to have an effective legal basis to process people’s information to train AI models.
Despite these challenges, last May Start notifying area users Regarding the upcoming privacy policy changes, the change notifies them that it will start using its comments, interactions, status updates, photos and subtitles for content in AI training. company debate This data processing is necessary for its AI model to reflect “various linguistic, geographical and cultural references to the European people.”
However, in June, Yuan was forced Put these plans on ice After review by the Irish Data Protection Commission (DPC), the EU’s Metadata Protection Commission’s lead data protector regulator has raised concerns about how Meta seeks users’ consent to process their data.
In short, Meta implements a heavy opt-out process (users must take action to prevent their data from becoming AI-training feed) rather than a simple choice – the company relies on GDPR legal basis called a “legal interest” and claims its actions are in line with the law. However, DPC disagrees and Meta is forced to rethink its approach.
The company confirmed to TechCrunch that the version of its AI Assistant launched in the EU has not been trained in data from local users, so why it said it would not notify users or otherwise seek their consent because it claims the technology has not been trained on its information.
“The model that powers these META AI capabilities is not trained on first-party data from EU users,” EMEA, Meta’s innovative communications manager, told TechCrunch.
Meta initially faced similar regulatory issues in the UK, as Brexit (because Brexit) is located outside the EU but still has a GDPR-based data protection regime.
Last summer, the UK’s Information Commissioner’s Office (ICO) Meta suspends its AI training program Worry about how it can help you grasp user data. But after the company Adjusted its opt-out process,Do Mild heavyMeta continues Launch Meta AI in the UK – No explicit objection from UK regulators, though ICO says it will “monitor the situation”.
When asked whether Meta’s AI efforts have not been trained in UK user data, a spokesperson directed TechCrunch to Announcement post Back in September, it said it would start training user content in the “next months”, indicating it is not ready for the public stage yet.
“Smart Chat”
Currently, the EU’s Meta AI will be limited to the company’s “smart chat functionality” calculations in six European languages: English, French, Spanish, Portuguese, German and Italian.
For now, the tool is basically a chatbot, which includes WhatsApp, Instagram, Messenger and Facebook itself, including WhatsApp, Instagram, Messenger and Facebook itself. It works by clicking a small blue circle icon to summon the assistant and asking any questions you might ask in search engines, such as how to perform tasks or find information about topics.
As mentioned above, the feature will also land in group chats, but this release is interleaved – starting with WhatsApp and then expanding to Messenger and Instagram direct messages “quickly.”
Users can call on assistants by typing “@metaai” and then asking a question, for example, where to have supper or top tourist attractions in a given city.
When announcing Meta AI “has a deep understanding of what you want,” Meta said that this is not referring to any personalized suggestions based on user data. It shows that marketing propositions are only in the context of easier, intuitive searching for content.
However, the company does emphasize that the launch represents its “first step” to bring more AI into Europe and plans to eventually plan to “find equality with the United States” over time. This could mean more quarrels between the monies – Always crucial European AI regulations and EU regulators.
TechCrunch contacted DPC to ask for its response to the EU’s announcement of Meta AI.
A spokesperson told us: “DPC, as the main oversight authority of META, has been conducting meta-AI checks with colleagues in the EU/EEA oversight authorities in recent months and we will continue to review when we sell to users in the coming weeks.”
