AI & Local News newsletter, issue 3
Hello, Welcome to our third issue of the AI & Local News newsletter. We’re heading into the year end with more news and a few not-to-miss events.
We’re extending the deadline for our AI & Local News Challenge through Sunday December 19. Join our December 13 Info Session to learn more. We’re looking forward to working with people that are building projects at the intersection of AI and news.
Hoping you had a nice Thanksgiving and best wishes for happy holidays to come.
Community & Project Lead, AI & Local News
NYC Media Lab
AP’s Local News AI initiative is nearing the all-50-states mark for news organizations that have completed the AI readiness scorecard. So far, 138 organizations have completed the challenge to understand where their operation sits for using tools and technology. The scorecard is still accepting newsroom insights, here’s where to start.
For those with questions about how AI can help with news and subscriptions, you can catch the AP team at these webinars:
Wednesday, Dec. 1 at 11 a.m. EST, JournalismAI Festival
How, when and why a news organization should consider AI
With Cynthia DuBose, Jay Allred and Pedro Burgos, with AP’s Aimee Rinehart as moderator. Watch the session video.
Thursday, Dec. 2 at 1 p.m. EST, Local Media Consortium
Fast Forward: Artificial Intelligence That Can Supercharge Your Newsroom With AP’s AI team, Jim Kennedy, Aimee Rinehart and Ernest Kung.
The first phase of our project with The Philadelphia Inquirer and The Texas Tribune is complete. Due to all the hard work done by our infrastructure engineer Sylvan Zheng, we’re now able to onboard additional newsrooms into our systems with ease. This is a major milestone for the scaling of our team’s work. In addition, we now have the necessary data in order to start building the machine learning model prototype for our recommendation system. We will be presenting the prototype to our partners in early December.
Our data scientist, Sam Petulla, ran an engaging workshop with members of The Texas Tribune team that will help us align machine learning outcomes with editorial values. The workshop invites participants to prioritize journalistic and fairness values in the context of fine-tuning the functionality of machine learning systems. We will be running a similar workshop with The Inquirer in January.
AI and Local News Challenge applications are open through December 19.
We’re looking for University, News Organization and Startup teams that want to apply artificial intelligence tools to the many opportunities for innovation and challenges in this crucial moment for journalism. Teams will attend presentations in cohort sessions, work with mentors that are experienced in journalism tech and innovation, and get feedback from people in the news industry.
Info Sessions Mon Dec 13 2pm ET (11am PT) RSVP
$7,500 per team
Teams receive mentorship from experienced media industry innovators
Program runs first half of 2022
Open to US based teams
100% remote program
Apply by Dec 19
PAI co-designed a panel with JournalismAI on Bridging AI Ethics and Journalistic Standards for the upcoming Journalism AI Festival
Thursday, 2 December, 4:15 PM – 5:00 PM GMT Attend or watch the session archive here
The Guardian’s data scientists have been working with other newsrooms to understand how AI can improve journalism — teaming up with Agence France-Presse (AFP) over the past six months to use machine learning to identify and source quotes from news articles. There are a number of benefits in using AI to extract quotes, from detecting and tabulating the diversity of sources quoted to fact-checking.
To train an AI model to identify quotes, AFP and Guardian journalists used two tools:
Spacy, which uses deep neural networks to process natural language;
Prodigy, an annotation tool to efficiently label data- such as text.
To train the model, AFP and the Guardian provided their style guides — along with a long list of regular expressions — as a baseline and then revised their list, adding new constructs and quotation styles. You can find a full description of how they identified quotes and speakers at project Github here. Despite some challenges in correctly parsing song lyrics and poems, the results look promising: the more the model learned, the better it got.
The project is part of the 2021 JournalismAI Collab Challenges, a global initiative that brings together media organizations to explore innovative solutions to improve journalism through the use of AI technologies.
Key strategies for publishers to leverage AI to improve their processes, free up time for their staff from mundane tasks and improve the quality of their journalism.
A new guide for news organizations to learn how AI can support their work. Produced by JournalismAI, a project of Polis–the journalism think-tank at the London School of Economics and Political Science (LSE)–supported by the Google News Initiative.
Date: November 29-December 3
The second edition of the JournalismAI Festival is being held between 29 November and 3 December. The 5-day program features speakers from newsrooms, media companies, and research institutions from around the world to talk about how journalism is evolving and the role of AI technology. Watch the live stream or session archive.
Wednesday, Dec. 1 at 11 am ET, JournalismAI Festival
How, when and why a news organization should consider AI
With Cynthia DuBose, Jay Allred and Pedro Burgos, with AP’s Aimee Rinehart as moderator.
Thursday, Dec 2 at 11:15 am ET Bridging AI Ethics and Journalistic Standards
*Note: We have a few free All-Access Passes available from the NYCEDC for NYC-area entrepreneurs and students interested in AI, particularly those from communities that are traditionally underrepresented in tech. Reach out to Matt MacVey (firstname.lastname@example.org) if you’re interested in a pass.
The Microsoft-backed artificial intelligence research firm Open AI just made its language supercomputer GPT-3 open to any eligible developer. GPT-3 can process 175 billion parameters and is used by apps like language-generating program Sudowrite. The Washington Post decided to test it, rewriting one of the most celebrated articles of the 20th Century: Gay Talese’s “Frank Sinatra Has a Cold.” Talese compared the two texts and determined that the original and the computer-generated were “both good.”
The consulting firm Infosys is partnering with Bloomberg Media to create the Bloomberg Digital Economy Index. Infosys will use analytics and AI to integrate the Bloomberg Index experience on Bloomberg Media's digital platforms and Bloomberg TV. The new tool will launch in 2022.
The Aspen Tech Policy Hub is accepting applications for proposals that implement expert recommendations from the Aspen Institute’s Commission on Information Disorder to alleviate the crisis of mis- and disinformation in America. Semi-finalists will be awarded $5,000, and a grand prize of $75,000 will be awarded to the winner. Applications are open for project proposals through January 10, 2022.
Kara Swisher talks to the “godfather of virtual reality,” Jaron Lanier, about how these technologies will continue to shape our lives, Facebook’s pivot to Meta and why Meta should be paying users for data.
The J Word, hosted by Robert Gutsche Jr, Professor of Critical Digital Media Practice at Lancaster University. This episode focuses on the potential role of artificial intelligence in journalism.
This documentary delves into the family-run newspaper keeping local journalism alive in Storm Lake, a small rural town in Iowa.
In our panel, practitioners developing and working with AI tools explored the opportunities and challenges that AI presents to news organizations.
Food for Thought
Can a Machine Learn Morality? Researchers at a Seattle A.I. Lab say they have built a system — name Delphi — that can make ethical judgments. But these systems could be as flawed and moral judgments are as complex as the people who create them. An interesting question posed by tools like Delphi is: Could AI ever write an op-ed?
THE AGE OF AI: And Our Human Future by Henry A. Kissinger, Eric Schmidt and Daniel Huttenlocher. Kevin Roose used Sudowrite, an A.I. writing program, to generate a review of the book — with mixed success.
“If we continually focused on a kind of Armageddon, doom-laden, dystopian future — as now, with climate breakdown — that’s where we’re going. We’ve got to be realistic. But we also have to be optimistic and hope that we can use all our ingenuity, plus the ingenuity of the tools we’re creating, to get us into a better place than we are now.”
- Jeanette Winterson talking about AI with Kara Swisher on Sway
The initiative is funded by Knight Foundation.
Newsletter produced by Angelo Paura