Meta bets big on wearable AI new AI agent as Zuckerberg pushes back against AI doom
Meta Expands Wearable AI Push With New Muse Agent and Next-Generation Glasses Meta is stepping up its investment in artificial intelligence with a major expansion of its wearable technology, while Chief Executive Mark Zuckerberg continues to argue that the industry should keep developing increasingly capable AI rather than collectively slowing its progress. At the company’s annual Connect conference in Menlo Park, California, Meta presented a series of new products and updates built around artificial intelligence. The announcements included expanded capabilities for the company’s Muse AI agent, new smart glasses, an audio-only version of its Ray-Ban glasses and a much lighter virtual reality device that is being redesigned around a glasses-style form. The announcements demonstrate how closely Meta is linking its artificial intelligence ambitions with hardware. Rather than limiting AI to smartphones and computers, the company is developing products intended to make its assistant available while people are moving around, working, travelling or carrying out everyday activities. Zuckerberg has previously said that he expects wearable AI eventually to become one of the main ways people interact with technology. The latest announcements provide a clearer picture of how Meta intends to pursue that objective, with glasses, AI agents and virtual reality products becoming increasingly connected. Meta Puts Muse at the Centre of Its AI Strategy One of the most important announcements at the conference concerned Muse, Meta’s recently introduced personal AI agent. Unlike a conventional chatbot that primarily responds to questions, Muse is designed to carry out tasks on a user’s behalf. Meta says Muse can help users manage projects and goals, make plans and perform practical tasks using connected applications and online services. The company has also designed the system to remember information that users have previously provided, allowing it to make suggestions and potentially take action without every instruction having to be repeated. During demonstrations at Connect, Meta showed Muse being used for activities including travel planning, making reservations and ordering groceries. The company has established partnerships with services including OpenTable, Spotify and Ticketmaster to help its AI agent interact with outside services. The intention is to move AI beyond simply producing an answer on a screen. Instead, Meta wants Muse to become something closer to a digital assistant that can take a user’s request and then complete some of the work required to achieve it. Meta introduced Muse earlier in September and says the system operates through a dedicated secure virtual machine. The company says this environment is designed to isolate the agent and the user’s information while providing additional controls over the actions Muse can take. Privacy and Control Become More Important as AI Takes Action The move towards AI agents also creates a different set of challenges from those associated with conventional chatbots. If an AI system is simply answering questions, an incorrect response can generally be ignored or corrected. An agent that can send messages, make purchases or book services has the ability to create real-world consequences. Meta says Muse has therefore been designed with additional safeguards. The company says users determine which applications the agent can access and what level of permission it receives. It also says Muse can request confirmation before sensitive actions, such as sending an email or completing a purchase. Meta says users will also be able to see an audit trail showing what Muse has done and what it plans to do. Credentials such as passwords and payment information are intended to remain protected rather than being directly visible to the AI agent. The company is also planning another security system called Muse Confidential VM. Meta says this will provide additional encryption for the virtual machine, with a key controlled by the user rather than Meta. These measures are important to Meta’s wider attempt to persuade consumers that AI agents can be trusted with increasingly personal tasks. The more useful an agent becomes, the greater the amount of personal information and digital access it may potentially require. Muse Is Being Designed for Wearable Devices Meta’s longer-term plan is to make Muse available through its AI-powered glasses. This would allow people to interact with the agent without reaching for a phone or sitting in front of a computer. A wearer could potentially ask the assistant a question, request information about their surroundings or tell it to carry out a task using voice commands. Because the glasses can remain with the user throughout the day, Meta sees them as a natural platform for a personal AI assistant. Meta is also developing a small physical device called Muse Charm. Zuckerberg described the product as a keychain-style device containing the Muse experience. It is expected to arrive around the holiday period in December. The palm-sized device is intended to give people access to their personal AI without requiring a smartphone, computer or glasses. It represents another experiment by Meta in finding ways for people to maintain continuous access to an AI assistant. New Ray-Ban Meta Audio Glasses Arrive Without a Camera Meta is also expanding its smart glasses range with a new audio-focused model. The Ray-Ban Meta Audio glasses are designed to provide access to music, telephone calls and Meta’s AI assistant without including a screen or video-recording system. The new glasses are due to go on sale in October with prices starting at $349. By removing the camera and display, Meta is offering a different type of wearable from its existing camera-equipped smart glasses. The decision is significant because privacy has been one of the recurring issues surrounding smart glasses. Camera-equipped glasses can raise concerns about people being recorded without their knowledge, particularly in public places or during conversations. An audio-only design avoids some of those concerns while retaining several of the features that make wearable technology useful. Users can listen to audio, make calls and communicate with an AI assistant without having to look at a screen. Meta Adds Hearing Features to Its Smart Glasses Meta is also expanding the practical uses of its existing AI glasses. The company announced a hearing
