OLMo's ImpACT licenses, an app to track fuel data, and more.

Subscribe

August 2023

AI2 Newsletter

Icons representing AI2's core values of impact, accountability, transparency, and collaboration. These include an abstract flower made of roulette curves, an arrow hitting a bullseye on a target, a stack of papers, and four hands touching.

Top Story

The AI2 ImpACT License Project

Regulations can help address the challenge of preventing potentially harmful uses of AI, but regulatory efforts take time to establish. In the meantime, what steps can individuals and institutions take to help ensure that their AI research and systems are responsible and ethical? AI2 introduces the ImpACT License Project, which adopts a risk-based rather than artifact-based approach that enables public reporting of violations and disclosure requirements around intended use and project inputs. 

Learn more ➞

A Computational Inflection for Scientific Discovery

"We need systems that can understand the context of scientists and help retrieve & synthesize knowledge for specific tasks." AI2 research scientist Tom Hope contributed an article for the CACM magazine alongside Semantic Scholar's Director Doug Downey and Chief Scientist/GM Dan Weld, former AI2 CEO Oren Etzioni, and Microsoft's Chief Scientific Officer Eric Horvitz about the need for a human-centric approach to AI.

Read the article ➞

AI2 Wildlands Presents the Fuels Data App

AI2's Wildlands team went to Turnbull National Wildlife Refuge to test an early version of their Fuels Data app, an application for mobile devices that helps wildland practitioners measure and track surface fuels in order to help manage fires. They're looking for more testers! Read the blog to contact the team and access the app link.

A photograph of two sets of hands. One is entering data into a document with a dry erase marker, and the other is pointing to something on an app on their phone.
Learn more ➞
A photograph of Pradeep with his wife, Modhurima, and daughter, Sanjana.

Featured AI2er: Pradeep Dasigi

What are the benchmarks Pradeep Dasigi built that are still being used today? Find the answer in our feature all about Pradeep!

Learn more ➞
An illustration of a mouse and an elephant on a see-saw, with the mouse weighing down the see-saw and the elephant up in the air.

I2D2: Smaller Is Better

In the world of language models, it is commonly believed that bigger is better—but the I2D2 model is here to challenge that!

Learn more ➞

More from us

➞ Because Language Podcast: A.I. Hype Hosedown

➞ Machine Learning in Climate Action

➞ Wikipedia’s Moment of Truth

➞ LLM-Blender: A Simple Ensemble Learning Framework for LLMs

learn more about the allen institute of artificial intelligence

Work with us

AI2 Newsletter archive

Twitter
LinkedIn
YouTube
Website

Semantic Scholar, 2157 N Northlake Way, Suite 110, Seattle, Washington 98103