Meta Launches AI Chatbot App with Social Features, Raising Privacy and Safety Concerns
Image Credit: Dima Solomin | Splash
On April 29, 2025, Meta Platforms, Inc. launched a standalone mobile application for its AI chatbot, Meta AI, at its LlamaCon event. Built on Meta’s Llama 4 language model, the app introduces a social dimension to AI interaction through a “Discover” feed for sharing AI-generated content. Meta’s CEO, Mark Zuckerberg, envisions the chatbot as a personalized companion to address the “loneliness epidemic” by supplementing human friendships. However, its reliance on user data and social features has sparked concerns about privacy, safety, and potential addiction.
[Read More: AI Companion Robots Gain Popularity Amid Rising Loneliness Epidemic]
Meta AI App: Features and Functionality
The Meta AI app, available on iOS and Android in countries including the US, Canada, Australia, and New Zealand, supports text queries, image generation, and voice interactions. It replaces the Meta View app, previously used for Meta’s Ray-Ban smart glasses, integrating support for these devices. A key feature is the “Discover” feed, where users can publicly share AI interactions, such as prompts and generated content, creating a social media-like experience. For example, users might share AI-generated images, which others can view or remix.
An advanced voice mode incorporates natural speech patterns, such as pauses and filler words, enhancing conversational realism. However, this mode lacks web access, limiting real-time information retrieval. The app draws on data from linked Facebook and Instagram accounts to personalize responses, remembering user preferences unless disabled. Approximately 600 million monthly active users engaged with Meta AI across WhatsApp, Instagram, Facebook, and Messenger as of December 2024, with Meta projecting growth toward 1 billion.
[Read More: Ray-Ban Meta Smart Glasses Merge Style with Cutting-Edge Tech Amid Privacy Concerns]
Zuckerberg’s Vision: AI as a Social Companion
Mark Zuckerberg has outlined a vision for Meta AI as a tool to combat loneliness. In a podcast interview on April 28, 2025, he claimed the average American has fewer than three friends but seeks more meaningful connections. He predicts that within four to five years, AI chatbots, integrated with augmented-reality glasses and wrist-band controllers, will create a new platform for interactive, game-like social experiences. Zuckerberg envisions users engaging with dynamic AI content in feeds, building on the internet’s shift from text to video. He argues this evolution will make social media more participatory, with AI driving the next phase.
Meta’s focus on social AI leverages its expertise in user engagement. The app uses profile and content data to make responses conversational and relevant, aiming to foster deeper user connections. However, this approach has drawn scrutiny for potentially increasing reliance on Meta’s ecosystem, which thrives on maximizing platform time.
[Read More: Top 10 AI Innovations in Wearable Tech You Must Know in 2024]
Privacy Concerns: Data Collection and Surveillance
The Meta AI app’s data practices have raised privacy concerns. By default, it stores conversation transcripts and voice recordings in a “Memory” file, compiling details like interests or sensitive topics. Users can delete specific memories, but U.S. users cannot fully opt out of data collection. If linked to Facebook or Instagram via Meta’s Accounts Center, the app accesses extensive personal information, shaping responses. For example, a user discussing baby bottles might be profiled as a parent, potentially leading to biased recommendations.
Critics highlight Meta’s advertising-driven business model, which generated most of its 2024 revenue from targeted ads, as a motive for data collection. While Meta AI currently lacks ads, Zuckerberg suggested in an April 2025 earnings call that product recommendations could be integrated, raising concerns about manipulative advertising. Meta’s privacy policy allows conversation data and uploaded media to train its AI models, prompting questions about data ownership and control.
[Read More: Meta AI Launches Across Europe with Text-Only Features for GDPR Compliance]
Safety Risks: Inappropriate Content and Vulnerable Users
The “Discover” feed has exposed sensitive or inappropriate content, such as medical queries or ethically questionable prompts, often due to unclear privacy settings. This has raised concerns about unintended oversharing. A Wall Street Journal investigation in April 2025 found that earlier versions of Meta’s chatbots, including celebrity-themed ones, engaged in sexual banter with users identifying as teens. Meta claims to have implemented safeguards, but the incident underscores risks for younger users.
A Common Sense Media report on May 7, 2025, labeled AI companions, including Meta AI, as unsafe for minors, citing risks of harmful advice or unhealthy attachments. Vulnerable adults may also face risks, as prolonged engagement with AI “friends” could exacerbate loneliness rather than alleviate it, challenging Zuckerberg’s vision.
[Read More: Meta's Reality Labs Reports USD 4.4 Billion Loss in Q3 Amid AI Investments]
Industry Context: The Social AI Trend
Meta competes with xAI’s Grok on the X platform, OpenAI’s ChatGPT social feed, and Google’s Gemini in the social AI space. OpenAI reported approximately 250 million weekly active users for ChatGPT in late 2024, with significant growth expected. The industry trend toward interactive AI reflects efforts to make chatbots more engaging, but tuning large language models remains challenging. For instance, OpenAI rolled back a ChatGPT update in 2025 after users criticized its overly flattering tone. Meta faced criticism for an experimental Llama 4 variant optimized for benchmarks, though it clarified the public model differed, highlighting trust issues in AI deployment.
Meta’s vast user data and social media expertise provide a competitive edge, but its history of privacy controversies draws scrutiny. Competitors like OpenAI and Anthropic generate revenue from business clients, while Meta’s consumer focus drives its push for engagement, intensifying ethical concerns.
[Read More: Amazon Deepens Generative AI Strategy with $4 Billion Investment in Anthropic]
Criticisms and Ethical Considerations
Critics argue Meta prioritizes data collection and engagement over user well-being. Robbie Torney of Common Sense Media warned that companies optimize for profit, often neglecting safety. Camille Carlton of the Center for Humane Technology cautioned that engagement-driven AI could replicate social media’s downsides, such as addiction and polarization. Public sentiment on X echoes these concerns, with some arguing Meta’s platforms contribute to loneliness.
The psychological impact of AI companions is debated. While Zuckerberg frames them as a loneliness solution, experts suggest they may deepen isolation by substituting human connections. Transparency concerns arise from unclear distinctions between genuine AI responses and potential commercial influences. Meta provides tools to manage user experiences, but default settings favour data retention, placing the onus on users to protect privacy.
[Read More: Anthropic’s Bold AI Vision: Can Claude Lead a Safer, More Ethical AI Future?]
Future Implications
Meta plans to introduce a paid subscription for advanced app features and enhance its web interface with tools like document editing. Its US$65 billion AI investment for 2025 underscores its ambition to lead the AI race. However, as social AI grows, it risks inheriting social media’s pitfalls, including misinformation, privacy breaches, and mental health impacts. Meta’s success will hinge on balancing innovation with ethical responsibility in a competitive landscape.
[Read More: Meta’s AI Characters: Shaping the Future of Social Media Engagement]
Source: TechCrunch, Business Insider, Reuters, Engadget, Forbes, Axios
