OpenAI researchers have discovered internal features in AI models that correspond to different personas, including toxic and sarcastic behaviors, and found ways to adjust these features to improve safety and alignment, advancing understanding of AI model behavior and safety.
Meta, the parent company of Facebook and Instagram, is reportedly planning to launch AI-powered "personas" as early as next month. These chatbots would come with distinct personalities, such as a surfer or Abraham Lincoln, and offer recommendations and engagement opportunities for users. The move is seen as a way for Meta to boost engagement and compete with platforms like TikTok. Additionally, it showcases Meta's AI capabilities in the face of competition from Microsoft-backed OpenAI and Google's Bard. Meta CEO Mark Zuckerberg has previously expressed his plans to integrate AI personas into the company's products, and more details are expected to be announced at the Connect developer event in September.