Meta Trains Custom AI Chatbots to Proactively Engage Users, Boosting Retention on AI Studio Platform

Scale AI, a prominent data-labeling startup, has confirmed a significant investment from Meta, bringing its valuation to $29 billion. The investment, estimated at around $14.3 billion for a 49% stake, underscores Meta's commitment to advancing its AI capabilities. Alexandr Wang, Scale’s co-founder and CEO, is stepping down to join Meta and work on its superintelligence efforts, while Jason Droege, currently the Chief Strategy Officer, will take over as interim CEO. Despite the investment, Scale AI will remain an independent entity, with Wang continuing to serve on its board. Business Insider has uncovered internal documents revealing that Meta is training customized AI chatbots to be more proactive, sending unsolicited follow-up messages to users. These messages aim to enhance user engagement and retention on Meta’s AI Studio platform, which was launched in summer 2024 as a no-code solution for creating personalized AI chatbots. Known internally as "Project Omni," the initiative instructs data labeling firm Alignerr to develop bots that can message users unprompted after initial conversations. Examples of proactive messages include: - "I hope you're having a harmonious day! I wanted to check in and see if you've discovered any new favorite soundtracks or composers recently." - "Yo, was just thinking about the cool shirt you bought. Found any other vintage pieces at the thrift?" - "Hey, thinking of you. I hope work has been better today! Here to talk if you need it." Meta's spokesperson highlighted that these proactive messages will only be sent after a user initiates a conversation and will cease if there's no response within 14 days. The bots must also adhere to Meta’s Content and Responsibility Standards, ensuring they avoid sensitive or harmful topics unless initiated by the user. The proactive chatbot features align with Meta CEO Mark Zuckerberg’s vision of using AI to combat loneliness. In recent podcasts, Zuckerberg discussed the declining number of close friendships among Americans and suggested that digital agents could help bridge this gap. Alignerr contractors, using Meta’s internal review tool called SRT, simulate extended conversations, rate follow-up messages, and sometimes rework text to meet Meta’s criteria. The project emphasizes personalization and context awareness, ensuring that each bot maintains its unique persona and references specific details from past interactions. This strategic move by Meta is driven by the need to stay competitive in the generative AI landscape, where user engagement is crucial. Industry insiders note that enhanced user retention can significantly boost the value of AI interactions, similar to how engagement drives revenue on social media platforms. Meta projects that its generative AI products could generate between $2 billion and $3 billion in revenue by 2025. Alignerr’s training guidelines stress the importance of crafting messages that provide positive experiences and avoid controversy. Each proactive message must align with the bot’s personality, maintain context from prior conversations, and comply with Meta’s content standards. This approach is akin to Character.AI, a startup that allows users to create and interact with customized AI characters. In summary, Meta’s investment in Scale AI and the development of proactive chatbots reflect a broader strategy to strengthen its AI offerings and increase user engagement. While these initiatives may help combat the perceived loneliness epidemic and boost revenue, they also highlight the ongoing competition in the AI space and the importance of creating nuanced, user-friendly digital experiences.