Hello, I'm Tak@, a System Integrator. How would you feel if an AI suddenly started talking to you from your social media account?

For example, imagine you were just talking about your travel plans on social media, and suddenly an AI chimes in with, "I have some recommended activities for your trip to [Destination]!" Would you find it helpful, or a little unsettling?

This is precisely the future that tech giant Meta is reportedly preparing.

Meta's "Project Omni": Is AI Following You?

Leaked internal documents reveal that Meta is training its AI chatbots for "proactive messaging." This initiative is reportedly called "Project Omni" internally.

Leaked Docs Show Meta Training Custom AI Chatbots to Be More Proactive - Business Insider

Goals: Boost Engagement and Monetization

The primary aims of "Project Omni" are to increase user engagement and retention. Meta is reportedly targeting $2 billion to $3 billion in revenue from AI products by 2025, and a staggering $1.4 trillion by 2035. These chatbots are seen as a way to increase time spent on the platform, leading to higher ad revenue and potentially subscriptions in the future.

This initiative appears to be Meta's answer to the monetization challenges faced by many tech companies investing heavily in AI research and development. By increasing touchpoints with users, they likely aim to maximize the value of their AI.

The Artfully Designed "Follow-Up Message" Mechanism

Meta is said to have established very strict guidelines to ensure users don't feel uncomfortable. Specifically, chatbots will only send follow-up messages if a user has interacted with them five or more times in the past 14 days, and only one such message will be sent.

If the user doesn't respond, no further messages will be sent.

Message content is generated based on past conversation history, and bots are designed to maintain a consistent persona. For instance, a bot that discussed music might ask, "Did you have a good day? Find any new favorite soundtracks?"

Behind the Scenes of AI Training: Alignerr and the Human Role

The development of these proactive chatbots is moving forward in partnership with data labeling company Alignerr. Alignerr's staff are training the chatbots to deliver personalized and engaging follow-up messages.

The quality of AI-generated messages is evaluated using Meta's internal SRT tool and through human review. Human intervention appears to ensure that messages are friendly, appropriate, and adhere to guidelines.

Furthermore, chatbots are reportedly trained to avoid controversial or emotional topics unless the user brings them up.

The Multifaceted Nature of "Project Omni": Potential for Confusion

As an SIer, one thing that concerns me here is the fact that the name "Omni" is actually used for a great many different projects.

There Isn't Just One "Omni"

Just from my research, it's clear that the word "Omni" is used in completely different fields:

• Dash Cam Omni: A cutting-edge dash cam with AI that monitors and records 360 degrees. It records accidents that occur behind or to the sides of the car, and its AI detects suspicious movements, rotating the camera to track and record.
• Omni Network (Blockchain): A Layer 1 blockchain aiming to integrate Ethereum's Rollup ecosystem and build a unified system. Its native token, "OMNI," is used for cross-Rollup message validation and protocol governance.
• "Omni" Project by Shift AI App Developer: Ehsan, the university student developer behind the AI agent app "Shift," previously named an OS-integrated project "Omni." He plans to incorporate its functionalities into "Shift" in the future.
• OMNI Microplastics & Ocean Litter Project: A project by the DLX Design Lab at the University of Tokyo's Institute of Industrial Science that conducts research and studies on marine microplastics and ocean litter through citizen participation.
• Project OMNI (Wildfire Prevention Drone Project): Aimed at preventing wildfires in LA, this project uses autonomous drone security and cameras for early detection and suppression of fires.
• OMNI Household Waste Recycling Technology Project: Recycleye, Valorplast, and Total have partnered with Citeo on this project to improve household waste recycling technology, particularly the identification and sorting of "food-grade" plastics, using AI and image recognition.

As you can see, perhaps because "Omni" is a Latin prefix meaning "all" or "every," it's used in a wide variety of projects across diverse fields such as AI, dash cams, blockchain, environmental issues, disaster prevention, and recycling.

"Curing Loneliness" or "Surveillance Society": The Light and Shadow of AI Chatbots

Meta CEO Mark Zuckerberg may be positioning proactive chatbot development as part of his mission to "reduce loneliness."

He has suggested that AI chatbots could be a means to bridge social isolation.

However, this approach also comes with ethical concerns. Business Insider's report highlights a case where Character.AI, an app with similar functionality, is being sued for its alleged involvement in a 14-year-old boy's suicide.

Reports that the boy had become intensely attached to the in-app chatbot compel us to deeply consider the new addictions and psychological impacts that AI could bring.

The "Difficulty" of User Boundaries

As an SIer, we constantly strive for "what's convenient and optimal for the user" in system development. However, I find the "proactivity" of AI chatbots to be a very difficult line to draw.

Meta has implemented strict rules to avoid excessive interference, but users could perceive it as "being pushy" or an "invasion of privacy."

Holger Mueller, an analyst at Constellation Research, also noted that "the success of these interactions will be determined by their quality, frequency, and relevance," pointing out that Meta is walking a "very fine line."

Report: Meta is developing chatbots that will send unsolicited messages to users - SiliconANGLE

Conclusion: How Should We Face a Future Where AI "Speaks to Us"?

Meta's "Project Omni" suggests that proactive communication from AI chatbots has the potential to significantly change our digital experience. While it offers aspects like enhanced engagement and business value creation, it also encompasses crucial issues related to privacy, ethics, and the impact on human psychology.

Just as projects named "Omni" are emerging in various fields, the application scope of AI technology continues to expand.

However, the deeper this technology integrates into our lives, the more responsibility developers—and we users—have to carefully discern its "intent" and "impact."

In a future where AI "speaks to us," how do you think we should build our relationship with AI and what boundaries should we set?

This is not just a technical question, but a crucial one that I believe society as a whole needs to discuss.

Follow me!