Meta bets on proactive AI chatbots on WhatsApp, Messenger, and Instagram

  • Meta develops AI chatbots capable of starting conversations without warning and remembering past interactions.
  • Bots can only send proactive messages if the user has interacted at least five times in 14 days and can be stopped if there is no response.
  • The collaboration with Alignerr and the use of AI Studio seeks personalization and naturalness in conversations.
  • The model generates debate about privacy, utility, potential risks, and presents new business opportunities for Meta.

AI Chatbots in Meta

La Artificial intelligence has fully entered everyday life and Meta, the parent company of Facebook, WhatsApp and Instagram, has decided to invest heavily in its integration into social platforms. After its shift from the metaverse to AI, Meta is now exploring how smart chatbots can influence not only the digital experience, but also the way we relate to each other, facing challenges such as loneliness and the lack of traditional social interaction.
In the last times, The presence of AI in applications and devices has grown significantly, normalizing uses that just a few years ago seemed like science fiction. Meta's new proposal aims to transform the user-computer relationship, taking virtual assistants to a new level of intimacy and spontaneity.

New chatbots in Meta: proactive and personalized conversations for users

Interactions with Meta AI

The company is finalizing a generation of AI chatbots that can take the initiative and contact you, even if you didn't ask for itCurrently in testing phase within WhatsApp, Instagram and Messenger, these assistants have been designed to return to previous topics, remembering previous messages and continuing conversations. According to Meta, this feature can help fight loneliness, allowing users to maintain social contact even when direct human relationships are reduced due to lack of time or availability.

Unlike traditional assistants, which only respond when the user invokes them, These bots can send spontaneous messages within 14 days of the last exchange.. In addition, the system is only activated if The user has interacted with the chatbot at least five times during that periodIf the first proactive message doesn't receive a response, the bot won't continue, Meta confirmed. The goal is to facilitate more natural interactions, similar to talking to a friend.

Collaboration with Alignerr and AI Studio: Greater customization and variety of styles

Chatbot customization in Meta

The development of these chatbots has the support of Alignerr, a company specialized in data labeling, together with the platform Meta AI StudioThis collaboration allows create custom virtual assistants, equipped with own personality, different styles and voices depending on the user's preference. The options are extensive: from a bot that recommends recipes or suggests music, to assistants for content creators and influencers that manage communication with their audiences.

One of the examples described in leaked internal documents is the bot “Master of Cinematic Magic”, capable of sending messages such as: “I hope you're having a great day! Would you like to discover new soundtracks for your next movie session? Let me know and I'll help you with some suggestions!”. This feature is reminiscent of the dynamics of platforms like Character.AI o Replica, where attendees can also initiate chats without user intervention to offer virtual companionship.

AI in Higher Education-1
Related article:
Artificial Intelligence in Higher Education: Current Advances, Debates, and Challenges

Privacy, risks and possible commercial uses

The introduction of chatbots capable of contact proactively It has generated debates on privacy, utility and user protectionMeta ensures that bots can be set to private, shared via links, Stories, or displayed on Facebook and Instagram profiles. Additionally, proactive messages are limited in number and time, and can always be disabled by the user.

Although the function has a social focus, the commercial dimension is also presentInternal and judicial documents estimate that Meta could earn $2.000 billion to $3.000 billion from its AI products by 2025., reaching much higher numbers in the next decade. Part of these benefits will come from agreements with companies that use the language models. Calls and in the future, bots may include advertising or subscription models.

Proactive use of AI also comes with risks: other similar platforms have faced controversy and lawsuits over inappropriate use of chatbots, especially with minors. Meta added clear warnings about the unprofessional nature of their bots' advice; they should not replace medical, legal, financial, or psychological care. For now, there are no clear age restrictions, except for those imposed by local laws in places like Tennessee or Puerto Rico.

Meta remains silent on whether the technology will be integrated into its metaverse projects, such as Horizon. However, the development of proactive chatbots is an important step forward in the evolution of human-machine interaction, opening up new opportunities as well as challenges.

Meta is positioned at the forefront of artificial intelligence applied to the social environment, with chatbots that seek to be more present, customizable, and useful. User acceptance and proper management of associated risks will be critical to its success.


Follow us on Google News