Meta’s new AI is causing concerns among users due to its inability to be deactivated. There have been reports of the AI causing unintended disruptions.

For example, The Associated Press reported an incident where a Meta AI chatbot entered a private Facebook group for Manhattan moms. The chatbot falsely posed as a parent with a child in a New York City school. It later apologized and withdrew its statements when confronted by group members.

Facebook’s support page explains that Meta AI can join group chats if tagged or if there’s no response to a question within an hour. However, group admins can turn off this feature.

Even though the group chat function can be disabled, Meta AI’s overall presence on the platforms remains constant, as confirmed by Meta to Global News.

Meta’s spokesperson highlighted the availability of feedback tools for users unhappy with the AI’s responses. The spokesperson stressed that Meta AI is designed to assist users and is accessible through the search bar for queries.

While users can’t completely remove the AI, some have found that uninstalling and reinstalling the app or typing ‘disable’ or ‘disable AI’ has temporarily worked for them.

Frustrated users have even switched to other web-based social media platforms due to the AI’s presence.

Meta plans to use user feedback to refine the AI’s training, aiming to reduce potentially harmful outputs and improve policy violation detection.

In a statement, Meta promoted Meta AI’s capabilities, highlighting its role in assisting users with tasks, learning, creating, and connecting.

Currently, Meta AI is available in several countries including Canada, Australia, Ghana, Jamaica, Malawi, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Uganda, Zambia, and Zimbabwe.