Meta (META.US) has launched the AI personal assistant Muse: a monthly fee of up to $100 aimed at the next generation of AI commercialization entry points.

On Tuesday, Meta (META.US) officially launched the AI personal assistant application Muse and made its first attempt to charge a subscription fee to some users. As the AI industry faces pressure regarding safety, privacy, and returns on significant capital expenditures, Meta is trying to push AI beyond chat tools to "personal assistants" that can perform tasks directly for users, aiming to make these assistants key products and revenue sources for the company's next phase. Muse, internally codenamed "Hatch," is based on the Muse Spark series of foundational models that Meta AI head Alexandr Wang has launched since April this year. Unlike traditional chatbots which mainly provide information and answer questions, Muse focuses on completing digital tasks for users, including making appointments, filling out online forms, and even monitoring home security cameras. Wang stated in an interview that Meta hopes Muse will feel simple, user-friendly, and easy to understand for the average user. While the interface seen by users is straightforward, Muse may be executing complex programming in the background, building integration between different services, and completing many intricate operations. Meta plans to offer a free version of Muse alongside subscription tiers at $20 and $100 per month, depending on users' usage levels. This pricing model signifies Meta's efforts to directly generate AI revenue from consumers. Zuckerberg has previously signaled to investors that more powerful personal AI assistants will be a significant new bet for the company. These assistants could not only help users organize emails but also actively seek product discounts and manage daily affairs. During Meta's earnings call in July, Zuckerberg mentioned that the new personal assistant would lay the groundwork for the company's next wave of products and revenue sources in the coming months and even years. This strategy also explains why Meta has continuously invested heavily in data centers, computational power, and other AI infrastructure in recent years. As capital expenditures keep rising, Wall Street increasingly wants to see these investments actual revenue, not just more advanced AI models. Muse is not the only recent attempt by Meta to commercialize AI. The company previously launched the Muse Code AI assistant for developers and is testing various AI subscription services, hoping to establish new income streams beyond traditional digital advertising. Not just chatting: capable of executing appointments, filling forms, and shopping tasks The biggest distinction of Muse is its "actionability." Users can authorize the assistant to carry out a range of real-world tasks, not just merely asking how to proceed. For example, Muse can help with scheduling appointments, filling out online forms, and executing tasks across different digital services. The Muse application also includes a news feed, which can aggregate updates from users' Facebook and Instagram accounts as well as articles and information from the internet, organized and summarized by AI. Muse will also be available on WhatsApp, although the WhatsApp version will primarily retain the core conversational and assistant functions and will not include the complete news feed and the Ideas tool that suggests ways to use the AI assistant found in the standalone Muse app. Meta stated that Muse will initially launch in the United States, accessible via iOS, Android, and a standalone website, with future plans to integrate it into Ray-Ban Meta smart glasses. Meta is exploring AI shopping commissions, and personal assistants may become a new commercial entry point In addition to subscription fees, Meta is investigating more significant potential business models, such as taking commissions on shopping transactions facilitated by AI assistants. Wang noted that since personal assistants can help consumers find products of interest and assist further in completing purchases, a commercial transaction model holds considerable potential for Muse. However, Meta has not yet determined specific pricing and revenue-sharing mechanisms. If this model comes to fruition, Muse's significance to Meta could extend far beyond the $20 or $100 monthly subscription revenue. Traditional internet platforms mainly guide consumers to businesses through search, social media feeds, and advertising, while AI assistants may soon directly participate in the entire process of "finding productscomparing pricesmaking choicescompleting transactions." This means that platforms controlling access to personal assistants have opportunities to further engage in the internet commercial transaction chain. For Meta, which currently relies heavily on digital advertising for revenue, this also represents a potential new growth direction. Concerns regarding AI assistant safety risks and user data being defaulted for model training However, as Muse launches, safety and privacy issues related to AI assistants are receiving increasingly strict scrutiny. Compared to general chatbots, personal assistants require higher permissions because they need access to user accounts, information, and third-party services, performing actions on behalf of users. Any model that comes under attack or misexecutes commands may lead to more severe consequences. Wang reported that Muse operates within an independent isolated environment on Meta's computational infrastructure and does not see users' real passwords or payment information, nor does it execute sensitive operations without first requesting user permission. It is noteworthy that Meta adopts an opt-out mechanism for Muse user data. This means that users need to actively choose to opt-out to prevent Meta from using their interactions with Muse to train AI models. Meta's Engineering VP David Singleton stated that if users do not opt-out, the company will remove "key personal identifiable information" from the dialogue and interaction data before using it to improve the model. Meta is also inviting third-party security researchers to identify vulnerabilities in Muse through a bug bounty program, offering rewards for qualifying discoveries. The company noted that before the official launch, Muse had been strengthened against security issues revealed through extensive internal testing, red teaming, and real-world adversarial environments. AI giants are competing for the entry point of personal assistants Metas release of Muse reflects a changing competitive landscape in the AI industry. With OpenClaw gaining rapid popularity, AI assistants capable of autonomously invoking models, tools, and software to complete complex tasks have become a new focal point of competition in the tech sector. Companies like OpenAI, Google, and startups like Town have launched related products or features in this area. Wang emphasized that compared to some competing products, Muse aims to lower the usage threshold of personal assistants, making them genuinely accessible to a broader consumer market. However, he also admitted that the industry as a whole is still in the very early stages of the personal assistant era. Furthermore, Metas launch of Muse faces additional pressures. The company recently agreed to pay nearly $17 billion to settle lawsuits related to potential harms associated with platforms like Facebook and Instagram brought by a coalition of state attorneys general, while still facing lawsuits associated with personal injury and school districts. Meanwhile, public skepticism regarding AI data center construction, energy consumption, and potential cybersecurity risks tied to AI is also escalating. Muse is not only a new AI product for Meta but also a crucial test of whether the company's massive AI investments can generate commercial returns. If Meta can get everyday consumers accustomed to letting AI handle appointments, shopping, information organization, and even household digital affairs, personal assistants could become another significant internet entry point for Meta, following the social media feeds. Subscription fees, AI shopping commissions, and future integrations with smart glasses could collectively form the core path for Metas next phase of AI commercialization.