Tag: Models

A new report reveals that ChatGPT, an AI chatbot popular since November 2022, struggles to mimic human workers in scientific research. Researchers have created a computer program to detect fake studies generated by the chatbot, which can still fool humans with its writing. Computer models are necessary to differentiate between real and AI-generated papers, as ChatGPT-created texts differ from human-written ones in various ways. However, this is just a proof of concept, and further studies are needed to create robust models that can maintain scientific integrity.

Prosus, the global consumer internet group, is expanding its generative AI solutions across its portfolio using OpenAI technology. With over 800 AI practitioners in 100+ countries, the Prosus AI Lab helps startups scale small data teams and access models and tools previously only available to larger tech companies. This expansion covers the Classifieds, Edtech, Food Delivery, and Payments & Fintech sectors, enabling state-of-the-art AI science and technology to drive innovation and business results.

Generative AI used by advertisers to depict BIPOC models is causing Digital Blackface concerns. Using AI can exclude rather than include diversification. Advertisers and agencies can combat this by asking the right questions & using ethical/legal frameworks. #DiversityInAdvertising

OpenAI improves AI by training models with process supervision, preventing errors. The method, tested in math problems, led to better performance and made AI more interpretable. There's a need to extend the study to ensure safety and ethical AI development.

🤖 Character.AI, an AI chatbot startup by Noam Shazeer and Daniel De Freitas, has had over 1.7 million installs in its first week. The app offers customizable AI companions with distinct personalities. #AI #Chatbot #Startup