Here’s how researchers are helping AIs get their facts straight

Having AI models say how confident they are in their answers could help minimize inaccurate responses. Just don’t be overconfident about their confidence scores.

Lu Wang, Associate Professor of Computer Science and Engineering, University of Michigan • conversation
Feb. 10, 2025 ~8 min

AI is transforming the search for new materials that can help create the technologies of the future

Discovering new materials could drive forward areas such as green technology and advanced electronics.

Domenico Vicinanza, Associate Professor of Intelligent Systems and Data Science, Anglia Ruskin University • conversation
Feb. 10, 2025 ~9 min


AI can boost economic growth, but it needs to be managed incredibly carefully

The risks of AI include displacement of jobs, which could counter any benefits.

Asieh Hosseini Tabaghdehi, Senior Lecturer in Strategy & Business Economy, Brunel Business School, Brunel University of London • conversation
Feb. 7, 2025 ~7 min

Even Bill Gates thinks AI is a little scary

Tech pioneer visits campus with his new memoir to discuss beauty of math, dropping out of College, founding Microsoft, value of curiosity

Harvard Gazette • harvard
Feb. 6, 2025 ~7 min

AI datasets have human values blind spots − new research

AI systems reflect human values. However, the human values embedded in AI are skewed to the utilitarian and away from the greater good.

Ike Obi, Ph.D. student in Computer and Information Technology, Purdue University • conversation
Feb. 6, 2025 ~5 min

Putting DeepSeek to the test: how its performance compares against other AI tools

This evaluation shows how competitive DeepSeek’s R1 chatbot is, beating OpenAI’s flagship models for performance as well as price.

Simon Thorne, Senior Lecturer in Computing and ​Information Systems, Cardiff Metropolitan University • conversation
Feb. 4, 2025 ~6 min

Government needs to show that its AI plan can be trusted to deal with serious risks when it comes to health data

The use of patient data has been controversial in the past.

Richard Milne, Head of Research and Dialogue, Wellcome Trust Sanger Institute • conversation
Feb. 3, 2025 ~7 min

DeepSeek claims to have cured AI’s environmental headache. The Jevons paradox suggests it might make things worse

In the 1860s, economist William Stanley Jevons said more efficient coal furnaces simply meant more coal was burned.

Peter Howson, Assistant Professor in International Development, Northumbria University, Newcastle • conversation
Jan. 31, 2025 ~7 min


How satellites and AI help fight wildfires today

Fire spotters used to watch with binoculars from forest towers. Now, technology can help forecast fire behavior, but human experience is still essential.

John W. Daily, Research Professor in Thermo Fluid Sciences, University of Colorado Boulder • conversation
Jan. 30, 2025 ~6 min

Why building big AIs costs billions – and how Chinese startup DeepSeek dramatically changed the calculus

A machine learning expert breaks down where the money goes in building big AIs, and how DeepSeek found ways to do it far more cheaply.

Ambuj Tewari, Professor of Statistics, University of Michigan • conversation
Jan. 29, 2025 ~7 min

/

61