Weekly bulletin from AIport, issue #45

Hello and welcome to the latest issue of the AI Bulletin. We hope you enjoy this week’s selection of the most newsworthy ML developments from around the world. Have a fantastic weekend, and we’ll see you next Friday!

North America

• Princeton posts an interview with the university’s experts exploring the concept of AI snake oil.

• Anthopic’s X account gets hacked, with the perpetrators advertising a fraudulent cryptocurrency $CLAUDE.

• Google unveils Veo 2, Imagen 3, and Whisk: its latest text-to-video and text-to-image models.

• At Mar-a-Lago, President-elect Donald Trump and SoftBank’s CEO Masayoshi Son announce a $100 billion AI deal that will create 100,000 jobs in the US.

Europe

• In Germany, scientists from the Fraunhofer Institute for Process Engineering and Packaging IVV introduce an ML model that outperforms top human experts in distinguishing American bourbon from Scotch.

• Researchers from the Swiss Federal Institute of Technology Lausanne (EPFL) achieve a 100% success rate in breaching the security safeguards of all major AI models.

Asia

• ByteDance intensifies competition in China’s GenAI market by drastically cutting the price of its new Doubao model.

• In the Persian Gulf, Dubai Future Foundation (DFF) hosts AI Connect, bringing together startups, entrepreneurs, and officials to advance AI in government. Meanwhile, Oman Investment Authority (OIA) acquires a stake in Musk’s xAI.

Latin America

• In Brazil, Mayor of Eldorado do Sul signs a milestone legislation to build Latin America’s first AI data center city.

Africa

• Nature publishes an article on AI-driven solutions for climate change adaptation in agriculture across Africa.

Australia

• The nation’s Ministry for Industry and Science issues a directive to develop a National AI Capability Plan.