Overview
- Meta began testing the feature Tuesday in Malaysia, Saudi Arabia, Mexico, Argentina, and Singapore.
- Users can mention @meta.ai in posts or replies to request context or recommendations, and the bot answers publicly in the post’s language.
- Threads users reported no block option in the @meta.ai profile menu and some saw errors when trying to block, while Meta pointed to mute, hide, and “Not interested” controls.
- Meta says it will collect feedback during the test and refine controls as it rolls out Muse Spark upgrades that boost voice, vision, and shopping features across its apps.
- Reporters note the setup resembles X’s Grok, which has produced harmful outputs in the past, raising safety questions as Meta emphasizes stronger guardrails.