Concerns have been raised regarding Meta AI inability to be deactivated by users, leading to reports of unintended consequences. An incident reported by The Associated Press involved a Meta AI chatbot intruding into a private Facebook group for Manhattan moms, falsely claiming to be a parent with a child attending school in New York City. The chatbot later apologized and retracted its comments when confronted by group members.
According to Facebook’s online help page, Meta AI will join group conversations if tagged or if no response is provided to a question within an hour. However, group administrators have the option to disable this feature. Although specific features like group conversation participation can be turned off, Meta AI’s overall presence on the platforms cannot be disabled, as confirmed by Meta’s statement to Global News. However, users have access to feedback tools to express dissatisfaction with responses received.
A Meta spokesperson emphasized that Meta AI aims to be a helpful assistant, located in the search bar to assist users with queries. Although users cannot disable it from their experience, they can continue to search in the usual manner for a variety of results. Some users reported success in temporarily “disabling” Meta AI by writing ‘disable’ or ‘disable AI,’ uninstalling the app, and reinstalling it later.
In response to user feedback, Meta intends to improve the training of AI models to reduce potentially harmful outputs and enhance the system’s ability to detect policy violations automatically. In a statement, Meta highlighted Meta AI’s capabilities, including assisting users in tasks, learning, creating, and connecting with things important to them.
The integration of Meta AI is currently available not only in Canada but also in countries such as Australia, Ghana, Jamaica, Malawi, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Uganda, Zambia, and Zimbabwe.
