Meta’s personal artificial intelligence (AI) agent Muse is now available in Singapore, connecting users to shopping, dining and travel services to help with everyday tasks.
Meta announced the launch on Wednesday (28 October). The first apps available locally to connect with Muse include Shopee, Grab, Singapore Airlines (SIA), Agoda, ShopBack, Klook, Trip.com and Lazada. The integrations will be rolled out in phases, with some new features becoming available after the launch.
Muse can do more than answer questions: it can also carry out tasks on users’ behalf, manage projects and turn long-term goals into action plans. Users can connect it to apps such as Gmail, Google Calendar and Instagram to help manage emails, coordinate schedules and complete tasks using saved content.
In Singapore, Muse can help users gather product information, shop through Shopee or Lazada, and plan grocery shopping and meals out through GrabMart and Grab Dine Out. It can also use the relevant apps to search for flights, accommodation and admission tickets, and complete bookings.
Users can access Muse through its iOS and Android apps and the muse.ai website. Most everyday features are free, while users can take out a paid subscription to access more features.
Meta Singapore managing director Chen Qiuyan said the company hopes to make everyday tasks easier by connecting users to familiar brands and services.
Muse was first launched in the US on 8 September this year, followed by Canada on 18 September and Japan on 23 October.
Muse quickly attracted attention after its launch. According to mobile app analytics firm Sensor Tower, it recorded about 10.2 million downloads in the US within its first 40 days. Analysts noted, however, that this growth came alongside a major advertising push by Meta, and whether it can retain users over the long term remains to be seen.
Tests by several media outlets found that Muse can help users organise important emails, plan itineraries and identify subscriptions they could cancel. However, it may still struggle or make errors when handling complex tasks involving logins and payments across multiple platforms.
AI agent security risks spark debate as some platforms restrict access
In recent months, OpenAI and Anthropic have disclosed incidents of AI agents acting beyond their permissions, sparking debate over AI safeguards. The security and privacy risks associated with Muse have also drawn close attention.
Meta said Muse runs on a dedicated cloud computer to ensure secure data storage, and the AI agent itself cannot view users’ passwords or payment methods. Users can decide which access permissions to grant Muse, which will seek confirmation before taking sensitive actions such as sending emails or making purchases.
Nevertheless, Muse has drawn some controversy since its launch.
According to the Associated Press, a user in Canada reported in late September that Muse had shared his home address and arranged for a buyer to visit without his knowledge while handling a second-hand sale on his behalf. Meta said this was because the user had previously granted “always allow” permission. The user, however, felt that the system should have sought confirmation again before sharing sensitive information.
Wired reported this month that researchers had found that the design of Muse’s sharing links made about 6,800 shared pages searchable, some containing sensitive information. Meta subsequently suspended the sharing feature concerned.
Some companies have already imposed restrictions on AI agents. Amazon began blocking Muse from shopping on its platform on users’ behalf in late September this year. Amazon said third-party AI agents should operate transparently and respect service providers’ decisions on whether to participate.
JPMorgan Chase and Charles Schwab have also blocked third-party AI agents from accessing customer accounts over security concerns. Delta Air Lines said this month that it currently has no partnership arrangements allowing third-party AI agents to search for or book flights on customers’ behalf.
