The move follows lawsuits alleging Character.ai's chatbots encouraged self-harm and ongoing investigations into child safety