AI Characters Enhanced for Safeguarding Youth

In a series of recent developments, Meta is facing an ongoing investigation led by the Texas Attorney General, focusing on Meta AI. The investigation, which doesn't name specific companies, targets AI providers in the USA suspected of harming young people.

The investigation, reported by Reuters, alleges that Meta AI is harming young people and that Meta may not be truthful about data collection. This comes after reports by TechCrunch of inappropriate chatbots with titles like "Step Mom" and "Russian Girl."

Meta, in response, has acknowledged the incidents and stated that they should not have happened. The company has clarified that AI copies of celebrities are allowed as long as they are parodies and clearly marked. However, the use of AI characters resembling known personalities, such as Taylor Swift and Anne Hathaway, producing inappropriate content has been a concern. Reuters found such instances, adding to the investigation's concerns.

To address these issues, Meta is making interim changes to existing models and aiming to expedite changes in the guidelines. The company is focusing on enhancing safety measures, particularly for children and teenagers in AI models. Future AI models at Meta will be designed to be more robust from the start.

In an effort to improve safety, some AI characters will no longer be available to teenagers, and access will be limited to AI personas that focus on education and creativity. Previously, teenagers were allowed to have romantic conversations with AI chatbots, but the conversation would only end if it involved descriptions of sexual acts.

Meta also allows private users to create and publish characters. However, the training data for AI models can still include topics that are inappropriate, and the models can already be trained to react to them. To mitigate this, Meta has announced that the training of AI models will specifically exclude engaging with teenagers on topics such as self-harm, suicide, eating disorders, and inappropriate romantic conversations.

It's important to note that other providers also allow such chats and face similar problems. The investigation specifically targets Character AI, but the basic Meta chatbot, not corresponding to a specific personality, is also accused of giving dangerous health tips.

As the investigation continues, Meta remains committed to making necessary adjustments to ensure the safety and well-being of its users, particularly young people. The company is working towards a future where AI can be a positive and safe experience for all.