Weekly bulletin from AIport, issue #28
Nvidia gets caught AI scraping, South Korea unveils its first open-source AI model, Aussie magazine receives flak for AI-generated content, and much more.
Hello and welcome to the latest issue of our AI bulletin. We hope you enjoy this week’s selection of the most interesting ML and data science happenings. Don’t forget to leave us a thumbs-up and see you next week!
North America
Hugging Face, a key AI global player, buys XetHub, a collaboration platform for developers, marking the company’s largest acquisition to date.
Google DeepMind team in California publishes an academic paper on the creation of a highly competent table-tennis-playing AI robot.
A new report by 404 Media, drawing on internal emails and Slack chats, claims Nvidia has been web-scraping vast quantities of YouTube videos illegally to train its AI models.
Asia
In Indonesia, fishermen are using the new AI-backed NN Marlin app, developed by the National Research and Innovation Agency (BRIN), to locate the densest fishing grounds.
In China, TikTok’s parent company ByteDance releases JimengAI, its first text-to-video AI generator, on the App Store. Meanwhile, another Chinese company, ZhipuAI, decides to open-source its text-to-video diffusion model, CogVideoX, accompanied by a research paper.
South Korea’s LG suddenly unveils the country’s first open-source AI model, EXAONE 3.0.
Europe
Several tech companies send a written plea to the European Commission, requesting that the compliance consultation on AI models under the EU AI Act be postponed. In case you missed it, be sure to check out our summary of the Act here.
Yoshua Bengio, the veteran computer scientist from Montreal known as a godfather of AI, joins the UK-based Safeguarded AI project, which aims to design a new type of ML system that would monitor AI-reliant solutions.
Latin America
Brazil’s Ministry of Science, Technology and Innovations posts the nation’s new AI Plan for 2024-2028 on its website.
Africa
The Institute for Security Studies (ISS), headquartered in Pretoria, publishes a thinkpiece assessing various AI tools capable of combating environmental crime in Africa, such as poaching and illegal mining.
Australia
The highly reputable Adelaide-based science magazine, Cosmos, is facing nationwide criticism from journalists and readers for publishing AI-generated articles.