AI Roundup: Latest Stories and Experiments

Keeping up with the rapidly evolving AI industry is challenging. Here’s a summary of recent stories in the world of machine learning:

OpenAI and Arizona State University Partnership

This week, OpenAI partnered with Arizona State University (ASU) to bring ChatGPT, OpenAI’s AI-powered chatbot, to the university. ASU will run an open challenge in February to invite faculty and staff to submit ideas for using ChatGPT. This collaboration highlights the changing attitudes toward AI in education, as schools reconsider the potential of GenAI tools for learning.

The Role of GenAI in Education

Despite concerns over cheating, there are potential benefits of GenAI in education. For instance, a tool like ChatGPT could assist students with homework assignments, math problems, or essay outlines. It also raises questions about the incentive structure in education, leading to discussions about how GenAI could be used to engage students in new ways.

Other Notable AI Stories

In addition to the OpenAI-ASU deal, Microsoft made its Reading Coach AI tool available for free, while EU regulators seek greater algorithmic transparency in music streaming platforms. Other stories include NASA’s self-assembling robotic structure, Samsung’s AI-powered Galaxy S24, DeepMind’s new AI system, and Microsoft’s Copilot. These developments hint at the diverse applications of AI across different industries.

Advances in Environmental Science and Natural Sciences

In environmental science, researchers at EPFL have developed a new AI training method that could revolutionize how AI programs are trained for research needs. Similarly, Los Alamos National Lab has introduced a new approach for creating imagery, while startups like Pano AI are applying AI for wildfire detection.

Meanwhile, Los Alamos is also working on a new AI model for estimating the decline of permafrost. Biologists are exploring AI tools for tracking wildlife, and Argonne NL researchers are using machine learning to identify the best molecules for storing and controlling hydrogen as fuel.

Challenges in AI Applications

Despite these advancements, a study found that machine learning models used to predict patient responses to treatments were only accurate within the sample group they were trained on, highlighting the need for thorough testing in diverse populations and applications.