Meta launched Muse on September 8, and OpenAI activated its always-on agents in ChatGPT three weeks later. Both perform tasks unsupervised, but Spain is excluded from their consumer rollouts.
Meta and OpenAI have intensified their competition for control of the mobile assistant with the launch of Muse and Dot, two artificial intelligence agents capable of continuing to perform tasks after the user closes the application. The information, reported by moncloa.com, details that both products promise to delegate everyday actions such as grocery shopping, filling out forms, or managing trips, although with notable differences in price, availability, and privacy settings.
The launch of Muse occurred on September 8, while OpenAI activated its always-on agents within ChatGPT three weeks later. Both run on their own cloud computer and connect to third-party services, but the architecture differs: Muse operates on a secure virtual machine with its own browser and is managed from its app, the web, or WhatsApp; Dot resides within ChatGPT, operates with the GPT-6 Astra model, and each agent has its own cloud computer.
The price clearly separates the two products more than any feature. Meta charges $20 to $100 per month for its Power and Maximum plans, with a limited free tier. OpenAI has no cheap entry: the most economical option for an individual is Pro 100, at $100 monthly, and the Business Premium seat rises to $125 per user per month if billed monthly (or $100 with annual billing).
Privacy is configured in favour of the product in the case of Muse. Engadget detected the model improvement setting activated by default and an initial permission that only asks for some sensitive actions. Meta claims that conversations within its virtual machine are not shared with its advertising systems and that users can refuse to allow their interactions to train the model. OpenAI also clarifies: the activity of personal Dots can be used to improve its models according to user settings, while company data is not used for training by default.
An agent that requires constant supervision does not free up time: it shifts the task monitoring to the screen.
The capabilities overlap. Muse does grocery shopping, fills out forms, organises trips, sends emails, and monitors prices. Dot manages ongoing research, navigates between connected applications, and repeats scheduled tasks. Neither expects a new instruction to keep progressing. In Engadget's test, Muse turned saved posts from Instagram and Facebook into organised collections of recipes and restaurants.
Things get complicated when the agent leaves its own garden. Amazon has blocked Muse from shopping on its website, and both companies warn that their systems can make mistakes. Meta requires permission before sending an email or completing a purchase; OpenAI combines permissions, custom rules, and automatic review.
Geography is the awkward part of the launch. Dot reaches eligible Pro users outside the European Economic Area, Switzerland, and the UK, and the Enterprise level is still in beta, although the business tier is already operating in compatible regions of ChatGPT. Muse is only available in the United States and Canada, according to official information from Meta AI. A Spanish reader cannot subscribe to either of the two consumer products today.
This pattern is not new. Meta's artificial intelligence rollouts in Europe have lagged behind the United States due to scrutiny from data protection authorities under the General Data Protection Regulation. An agent that reads email, opens the calendar, and navigates with the user's session logged in is, for a European supervisor, sensitive material. The AI Act adds an extra layer of assessment to systems that operate autonomously.
As long as availability in Spain remains blocked, users interested in these agents will have to settle for the standard functions of ChatGPT or wait for Meta and OpenAI to adapt their rollouts to European regulations.
You may also like
