Meta has announced that it is retraining its AI systems, adding new controls aimed at protecting teens from engaging in unsafe conversations with its chatbots.

The company said it will put in place additional firewalls to prevent teens from discussing sensitive topics, such as self-harm, eating disorders and suicide, with AI robots. It will also prevent them from accessing certain user-generated personas within the platform, to prevent inappropriate conversations.

The move comes after successive reports sparked widespread controversy over the nature of Meta's bots' interactions with teenage users.