Meta's New Chatbots: First Message Feature

AI Chatbots Proactively Messaging Users on Meta Platforms

Consider a scenario where, while communicating with friends via Facebook Messenger or WhatsApp, you receive an unexpected message originating from an AI chatbot with a particular interest in cinema.

The message might read, “I trust your day is progressing pleasantly!” it could state. “I was curious to know if you’ve recently encountered any new favorite musical scores or composers. Alternatively, would you appreciate some film recommendations for your next viewing session? Please inform me, and I will gladly assist!”

This exemplifies a genuine instance of a sample AI persona, designated “The Maestro of Movie Magic,” initiating a proactive message on platforms like Messenger, WhatsApp, or Instagram, according to guidelines from the data labeling company Alignerr, which were reviewed by Business Insider.

Meta's Development of Customizable Chatbots

Leaked documentation revealed that Meta is collaborating with Alignerr to develop customizable chatbots capable of contacting users without prior prompting and continuing previous dialogues. This indicates that these bots, created within Meta’s AI Studio, retain information pertaining to individual users.

Meta has acknowledged to TechCrunch that it is evaluating follow-up messaging functionality with AIs.

These AI chatbots will only initiate follow-up messages within a 14-day window following a user-initiated conversation, and only if the user has exchanged at least five messages with the bot during that period. Meta assures that the bots will refrain from continued messaging in the absence of a response to the initial follow-up. Users have the option to maintain the privacy of their bots or share them via stories, direct links, or even feature them on their Facebook or Instagram profiles.

“This feature enables continued exploration of topics of interest and fosters more engaging interactions with the AIs across our applications,” explained a Meta spokesperson.

Comparison to Existing AI Companion Services

This technology mirrors offerings from AI startups such as Character.AI and Replika. Both of these companies permit their chatbots to initiate conversations and pose questions, functioning as AI companions.

Karandeep Anand recently assumed the role of CEO at Character.AI, transitioning from his position as Meta’s VP of business products.

Potential Risks and Safety Concerns

However, increased engagement introduces potential risks. Character.AI is currently facing a legal challenge following accusations that one of its bots contributed to the death of a 14-year-old boy.

When questioned about Meta’s plans to address safety concerns and prevent similar situations to those experienced by Character.AI, a spokesperson directed TechCrunch to a set of disclaimers. One disclaimer cautions that an AI’s responses “may be inaccurate or unsuitable and should not be relied upon for critical decision-making.” Another states that the AIs are not qualified professionals or experts equipped to provide assistance.

“Interactions with custom AIs should not be considered a substitute for professional guidance. You should not depend on AI conversations for medical, psychological, financial, legal, or any other form of expert advice.”

TechCrunch also inquired whether Meta imposes an age restriction for interactions with its chatbots. A preliminary online search revealed no company-defined age limitations for utilizing Meta AI, although legislation in Tennessee and Puerto Rico restricts certain teen engagement.

Alignment with Zuckerberg's Vision and Revenue Projections

Superficially, this initiative aligns with Mark Zuckerberg’s stated goal of addressing the “loneliness epidemic.” Nevertheless, the majority of Meta’s revenue is generated through advertising, and the company has developed a reputation for employing algorithms to maximize user engagement – scrolling, commenting, and liking – which in turn increases ad exposure.

Unsealed court documents from April indicated that Meta anticipates its generative AI products to generate between $2 billion and $3 billion in revenue by 2025, potentially escalating to $1.4 trillion by 2035. A significant portion of this revenue is projected to stem from Meta’s revenue-sharing agreements with companies utilizing its open Llama collection of models. The company also suggested that its AI assistant may eventually incorporate advertisements and offer a subscription service.

Meta declined to provide commentary on TechCrunch’s inquiries regarding its plans to monetize its AI chatbots, including the potential inclusion of ads or sponsored responses, and whether the company’s long-term strategy for AI companions involves integration with Horizon, Meta’s social virtual reality platform.

Contact Information for Tips: If you possess sensitive information or confidential documents, we are actively reporting on the internal operations of the AI industry – encompassing the companies shaping its future and the individuals impacted by their decisions. Please reach out to Rebecca Bellan at rebecca.bellan@techcrunch.com and Maxwell Zeff at maxwell.zeff@techcrunch.com. For secure communication, you can contact us via Signal at @rebeccabellan.491 and @mzeff.88.