AI News – July 20, 2023

Emerging Alternatives to ChatGPT

TechXplore reports on the development of new AI models that are emerging as alternatives to OpenAI’s ChatGPT. These models, such as EleutherAI’s GPT-NeoX and Google’s MUM, are being designed to offer similar capabilities to ChatGPT but with their own unique features. GPT-NeoX, for instance, is being developed as an open-source project, making it accessible to a wider range of users and developers. Google’s MUM, on the other hand, is designed to understand and generate responses across multiple languages and tasks, making it a potentially more versatile tool. Read more: TechXplore

AI in Q&A Format

Another article from TechXplore discusses the use of AI in a question-and-answer format. This approach is being used to make AI more accessible and user-friendly, allowing users to interact with AI in a more conversational manner. This can be particularly useful in areas such as customer service, where AI can handle routine inquiries, freeing up human staff to deal with more complex issues. The article also highlights the importance of ensuring that AI models are trained on diverse and representative data to avoid bias in their responses. Read more: TechXplore

Changes in ChatGPT’s Behavior

VentureBeat reports on recent changes observed in the behavior of ChatGPT. AI researchers have noted that the model’s responses have become more cautious and less creative over time. This is likely due to updates aimed at making the AI safer and less prone to generating inappropriate or harmful content. However, some users have expressed concerns that these changes may limit the AI’s usefulness and versatility. Read more: VentureBeat

Meta’s New Art-Generating Model

TechCrunch reports on Meta’s new AI art-generating model, CM3Leon. This model is claimed to achieve state-of-the-art performance for text-to-image generation and is one of the first image generators capable of generating captions for images. Meta claims that CM3Leon requires five times less compute and a smaller training dataset than previous transformer-based methods. The model was trained on a dataset of millions of licensed images from Shutterstock. Read more: TechCrunch

Authors Urge AI Makers to Stop ‘Stealing’ Books

TechCrunch reports that over 8,500 authors have signed an open letter urging tech companies to stop using their writing for AI model training without permission or compensation. The authors argue that AI technologies mimic and regurgitate their language, stories, style, and ideas, which they believe constitutes theft. The letter calls on the companies to obtain permission for the use of copyrighted material, compensate writers fairly for the past and ongoing use of their works, and compensate writers for the use of their works in AI output. Read more: TechCrunch

Meta Releases Llama 2

Meta has announced the release of Llama 2, a new family of AI models designed to drive apps such as OpenAI’s ChatGPT, Bing Chat, and other modern chatbots. According to TechCrunch, Llama 2 improves significantly over the previous generation of Llama models. The new models come in two flavors, Llama 2 and Llama 2-Chat, the latter of which was fine-tuned for two-way conversations. Read more: TechCrunch

Apple Testing ChatGPT-Like AI Chatbot

TechCrunch reports that Apple is developing a ChatGPT-like AI chatbot. The tech giant has created a chatbot that some engineers are internally referring to as “Apple GPT.” The company has yet to determine a strategy for releasing the technology to consumers, but is reportedly aiming to make a significant AI-related announcement next year. The report also mentions that Apple is focused on trying to address potential privacy concerns related to artificial intelligence. Read more: TechCrunch

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top
Share via
Copy link