AI Weekly Digest Issue #16

Welcome back to AI Weekly Digest #16, your go-to resource for the most important artificial intelligence news and events from Week 31 of 2023.

As an AI enthusiast, staying informed about the rapid advancements in the field is crucial for both personal and professional growth. Our mission is to bring you a concise and comprehensive roundup of the latest breakthroughs, innovations, and discussions shaping the world of AI.

Aug 04, 2023

4D AI: The Next Frontier in Artificial Intelligence

Artificial intelligence is moving beyond the three dimensions of space and time to a fourth dimension: the dimension of data. This new 4D AI is able to process and understand data in a way that was not possible before, opening up new possibilities for AI applications in fields such as healthcare, finance, and transportation.

From techcrunch

Amazon’s Entertainment Business Bets on Generative AI to Innovate

Amazon’s entertainment business is exploring multiple generative AI initiatives, according to CEO Andy Jassy. In an interview with Variety, Jassy said that the company is working on AI-powered tools that could be used to create new content, improve existing content, and personalize the user experience.

From variety

Air Force’s AI-Powered Plane Takes to the Skies

The U.S. Air Force has successfully completed the first flight of a pilotless plane controlled entirely by artificial intelligence. The plane, a modified C-130 Hercules transport aircraft, took off and landed autonomously at Edwards Air Force Base in California. The flight is a major milestone in the development of AI-powered aircraft, and could pave the way for future applications in military and commercial aviation.

From stripes

Aug 03, 2023

OpenAI just released a bevy of updates for ChatGPT

1. Prompt examples: A blank page can be intimidating. At the beginning of a new chat, you’ll now see examples to help you get started.
2. Suggested replies: Go deeper with a click. ChatGPT now suggests relevant ways to continue your conversation.
3. GPT-4 by default, finally: When starting a new chat as a Plus user, ChatGPT will remember your previously selected model — no more defaulting back to GPT-3.5.
4. Upload multiple files: You can now ask ChatGPT to analyze data and generate insights across multiple files. This is available with the Code Interpreter beta for all Plus users.
5. Stay logged in: You’ll no longer be logged out every 2 weeks! When you do need to log in, you’ll be greeted with a much more welcoming page.
6. Keyboard shortcuts: Work faster with shortcuts, like ⌘ (Ctrl) + Shift + ; to copy last code block. Try ⌘ (Ctrl) + / to see the complete list.

From openai

Uber Developing Its Own AI Chatbot, CEO Reveals

Uber is working on building its own AI chatbot, disclosed CEO Dara Khosrowshahi this week. With generative AI gaining popularity, Uber aims to incorporate the technology into its platform. While Uber already uses AI for matching riders and other functions, its own conversational bot would be a new frontier.

From zdnet

Apple Stays Mum on AI Amid Big Tech Hype

Unlike peers, Apple made barely any mention of artificial intelligence on its latest earnings call. While Microsoft, Google, Meta, and Amazon touted AI efforts, Apple CEO Tim Cook said AI enables many Apple products but didn’t elaborate. When asked, Cook said Apple believes in AI as a “core fundamental technology” and has done “research across a wide range of AI technologies” for years.

From zdnet

AI.com Domain Bounces from ChatGPT to X.ai

The valuable AI.com domain has changed hands again between AI heavyweights. Earlier this year, AI.com redirected to ChatGPT’s site, assumedly purchased by OpenAI. Now it goes to X.ai, Elon Musk’s new AI research group.

From techcrunch

IBM and NASA Open Source Large Geospatial AI Model

IBM and Hugging Face have publicly released IBM’s watsonx.ai geospatial foundation model, built using NASA satellite data. It is the largest open source geospatial AI model on Hugging Face. The model aims to help scientists analyze large climate datasets faster. Built via a NASA partnership, it was trained on satellite images across the U.S. and can be fine-tuned for tasks like flood mapping. Making it openly available can accelerate innovation in climate science and earth observation.

From ibm

Alibaba Challenges Meta, OpenAI with Open-Source AI Model

Chinese tech giant Alibaba has open-sourced its large language model Tongyi Qianwen, allowing global developers to access the technology for building conversational and generative AI apps. This poses potential competition to Meta’s open-sourced Llama model and OpenAI’s ChatGPT.

By distributing Tongyi Qianwen openly, Alibaba aims to gain more users and establish itself as a leader in AI. The move supports Alibaba’s cloud computing business as demand grows for strong AI capabilities. Though Alibaba won’t earn licensing revenue, increased adoption of Tongyi Qianwen could make it an attractive model for cloud providers like Microsoft Azure.

Alibaba has already built internal apps with Tongyi Qianwen like image generator Tongyi Wanxiang. As Chinese firms seek to advance in AI, Alibaba’s release of Tongyi Qianwen signals its ambitions to compete with western tech giants in generative AI.

From cnbc

Aug 02, 2023

Meta Open-Sources AI Framework for Generating Audio

Meta has open-sourced AudioCraft, a new framework for using AI to generate high-quality sounds, music, and speech. AudioCraft includes three models: MusicGen for music, AudioGen for sounds, and EnCodec for compressing audio. MusicGen was previously released, but now its training code is available too, raising ethical concerns around deepfaking artists’ voices. AudioGen uses diffusion to generate realistic environmental sounds from text prompts. EnCodec aims to compress audio with higher fidelity. While Meta touts AudioCraft’s potential creative uses, it acknowledges risks like voice deepfakes. With MusicGen’s Western bias and lack of foreign language support, Meta says it plans to improve model capabilities and controls. By open-sourcing AudioCraft, Meta enables exploration of AI audio generation, though consequences remain uncertain. The release represents Meta’s ongoing efforts to advance generative models for music and audio.

From techcrunch

Google AI Search Gets a Facelift with More Video and Better Links

Google is updating its AI search to include more video and better links. The new search will use AI to generate summaries of videos, as well as to provide more relevant links to websites. This will make it easier for users to find the information they are looking for, whether it is in the form of text, video, or images.

From theverge

Aug 01, 2023

Nvidia’s AI Image Generator: A Floppy Disk-Sized Breakthrough

Nvidia has developed a new AI image generator that is small enough to fit on a floppy disk and can be trained in just four minutes. The generator is based on a new type of AI model called a diffusion model, which is able to create realistic images from a small amount of data.

From decrypt

YouTube Tests AI-Generated Video Summaries

YouTube is experimenting with using AI to auto-generate video summaries, aiming to provide viewers with quick overviews to help them decide if a video is right for them. The AI-generated summaries are limited to English-language videos and a small subset of viewers for now.

From techcrunch

Meta prepares AI-powered chatbots in attempt to retain users

Meta is reportedly preparing to launch AI-powered chatbots in an attempt to retain users. The chatbots would be designed to help users with tasks such as customer service, product recommendations, and social media moderation. Meta is hoping that the chatbots will make it easier for users to get the help they need and to stay engaged with the company’s products.

From yahoo

OpenAI has filed a new trademark application for “GPT-5”

From reddit

July 31, 2023

Google Assistant to Pivot to Generative AI

Google is reportedly pivoting Google Assistant to generative AI, which would allow the virtual assistant to better understand and respond to natural language. Generative AI models are trained on large datasets of text and code, which allows them to generate text, translate languages, write different kinds of creative content, and answer your questions in an informative way.

From techcrunch

GPT-3 Can Reason as well as a College Student

A team of psychologists at Stanford University has found that GPT-3, a large language model developed by OpenAI, can reason as well as a college student. The researchers tested GPT-3 on various reasoning tasks, including syllogistic reasoning, analogy reasoning, and causal reasoning. GPT-3 performed as well as or better than college students on most tasks.

From techxplore

ChatGPT app for Android is now available in all countries and regions

MIT’s PhotoGuard: An AI Model to Protect Against AI Image Manipulation

Researchers at MIT have developed a new AI model that can help to protect against AI image manipulation. The model, called PhotoGuard, uses a combination of machine learning and computer vision to identify and flag images that have been tampered with.

From mit

Complex Media and Sprite Launch AI Chatbot to Engage Gen Z

Complex Media has partnered with Sprite to debut an AI chatbot called SpriteBot. SpriteBot is designed to help users learn more about Sprite and its brand partners, and to engage in conversation about pop culture and music. SpriteBot is powered by IBM Watson AI, and it is available on Facebook Messenger and Instagram.

From adweek

Stack Overflow Announces OverflowAI

Stack Overflow has announced the launch of OverflowAI, a new AI-powered platform that aims to help developers find the answers they need. OverflowAI uses a combination of natural language processing and machine learning to understand the context of a developer’s question and to provide relevant answers from Stack Overflow’s vast knowledge base.

From stackoverflow

LinkedIn’s AI ‘Coach’ to Help Users Land Their Dream Jobs

LinkedIn is reportedly working on an AI-powered “coach” that will help users with their job applications. The coach would use AI to analyze a user’s resume and LinkedIn profile, and then provide personalized advice on how to improve their chances of getting hired. The coach would also be able to help users practice their interviewing skills.

From theverge

Every week, we’ll meticulously curate a selection of stories from top AI media outlets and distill them into a digestible format, ensuring you stay up-to-date with the latest developments without having to spend hours browsing the web. From groundbreaking research to real-world applications, ethical debates to policy implications, AI Weekly Digest will be your essential guide to navigating the ever-evolving landscape of artificial intelligence. Join us on this exciting journey as we explore the future of AI together, one week at a time.