AI & Local News newsletter, issue 9
It’s an exciting time for the AI & Local News initiative organizations with new cohorts at Associated Press and Brown Institute’s Local News Lab and a new article from Partnership on AI. Read more about the work of those organizations below. Also–outside of the initiative–take a look at the projects of the JournalismAI Fellows of 2022.
This summer has been a fruitful time for new research that can inform our work in support of innovation to sustain local news. Reports from Northwestern University’s Medill Local News Initiative (more info below) and Reuters Institute for the Study of Journalism (27% of US news subscribers subscribe to a local news source) highlight the challenges and opportunities.
Finally, I’ve been fascinated by the new AI models–like DALL•E 2 from OpenAI–that generate digital images from text prompts. (Read about how OpenAI is trying to mitigate some of the risks and limitations of DALL•E 2) I’m curious if there are any applications for these models in the journalism world. Have you tried them, or seen any interesting experiments? Let me know!
Matt MacVey Community & Project Lead, AI & Local News NYC Media Lab firstname.lastname@example.org
AI & Local News Initiative News
Associated Press: “Associated Press’ Local News AI initiative will consult on projects that stretch from Alaska to Puerto Rico on technology that involve automatic translation, public records delivery into CMSs, audience question bots and more. AP will dig into project blueprints with KTUU-TV in Alaska, KYMA-TV in Arizona, Hearst Newspapers in California, Center for Public Integrity in Washington, D.C., in Iowa the Ottumwa Courier and Lee Enterprises, in Michigan WUOM-FM and the Detroit Free Press, The Haitian Times in New York, Forum Communications Company in North Dakota, in Ohio Richland Source and Lima News, El Vocero in Puerto Rico, The Post and Courier in South Carolina and KSAT-TV in Texas. Funded by the John S. and James L. Knight Foundation.”
Brown Institute’s Local News Lab: “In June, the Local News Lab launched its second cohort under the AI in Local News initiative. The lab’s first cohort with The Philadelphia Inquirer and Texas Tribune was a rewarding and successful project. We came away having learned a lot about newsroom collaborations and a strong recommendation system that performed well in A/B tests: our recommendations boosted reader engagement (as measured by clickthrough rate, or CTR) by 10% on the Tribune and 16% on The Inquirer.
Our second cohort consists of four newsroom partners in AfroLA, Dallas Free Press, Open Vallejo, and The 19th, this collaborative project has kicked off with early product discovery workshops. We are looking forward to seeing where this new group of partners - and the needs they have in common - takes the project next. The lab has also welcomed two new team members: Raaid Arshad (Lead Engineer) and Duy Nguyen (Machine Learning Engineer) have joined the lab for the duration of this cohort project. Both Raaid and Duy became quickly acquainted with our infrastructure and first project and are contributing to our product planning process with their deep technical expertise. We’re grateful for this new and brilliant iteration of our team.”
A few notable follow ups from the AI & Local News Challenge Cohort: Mike Stucka is one of the JournalismAI Fellows of 2022, Jessica Davis has joined the Oxford Climate Journalism Network cohort and Swapneel Mehta was accepted into the NYU Tech Venture Workshop.”
If we want to use AI to support local news, how can we make sure this technology’s business advantages don’t come at the expense of ethics or the societal benefits newsrooms provide? And, at a fundamental level, what should AI ethics for local news look like?
The Partnership on AI (PAI) believes that newsrooms using AI must consider AI ethics with the same rigor and sensitivity that they already apply to journalistic practices. Notably, such AI ethics should not be thought of as completely distinct from journalistic ethics, but rather an extension of them.”
Artificial intelligence is a more and more central part of newsrooms: without AI and algorithms, investigations like the ‘Pandora Papers’ would not have been possible. However in a recent article, the Oxford Internet Institute doctoral candidate Felix M. Simon explains that most AI tools, services and infrastructures, are concentrated in the hands of the well-known big tech names such as Microsoft, Amazon Web Services (AWS) and Google.
According to Simon, “the roll-out of AI in the news risks shifting even more control to major platform companies, making the news industry even more dependent on them than they are already.” Before the rise of AI big tech, platforms exerted their control over media via advertising and distribution channels, as well as by funding journalism projects. Now with AI they add new levels of control: production and connection, forcing media outlets to rely on their hardware, software, data, and expertise.
Here’s the full paper in Digital Journalism: ‘Uneasy Bedfellows: AI in the News, Platform Companies and the Issue of Journalistic Autonomy’
Oxford Internet Institute
According to Outlier founder and editor-in-chief Sarah Alvarez, solving the crisis of local news is not that complicated: what local news outlets have to do is to better understand the needs of the communities, and offer them the information ecosystems they deserve.
Outlier is a local media outlet based in Detroit with a clear mission: rebuild local journalism with a more in-depth approach to engaging communities and their needs.
As Alvarez wrote in a recent piece about Outlier: “What should be beyond argument is that it is time to replace local ‘news’ with more useful and inclusive visions of how reporters and newsrooms can serve their communities. It has to go far beyond just covering breaking news.”
Reynolds Journalism Institute
Dates: October 13-14 Location: Hybrid. Boston, MA and virtual. The annual Conference for Truth and Trust Online is a forum for academia, industry, non-profit organizations, and other stakeholders to discuss the problems facing (social) media platforms and technical solutions to understand and address them.
Northwestern University’s Medill Local News Initiative released their analysis of data and information on more than 8,000 newspapers and digital sites. The report highlights that local newspapers–especially weeklies–continue to close and be purchased by conglomerates. Digital first local news sites are growing, but not at a pace to offset the coverage lost in many communities. Read the report to understand the challenges facing local news and to inform appropriate solutions.
Northwestern University’s Medill Local News Initiative
New research suggests that diversity is an essential parameter for making AI a better team player: MIT Lincoln Laboratory researchers have found that training an AI model with "diverse" teammates improves its ability to collaborate with other AI in card game Hanabi. At the same time, both Facebook and Google’s DeepMind have recently published studies showing that introducing diversity into training improves outcomes in human-AI collaborative games.
MIT Lincoln Laboratory
The Local Media Foundation (LMF) just released a Knight-funded project report about a year-long project with four local media outlets to understand if a new content management system for digital sites delivered enough to justify the investment. The answer is complicated: a new CMS will not immediately ease difficulties faced by media outlets. However, it will “provide demonstrable enhancements to user experience, lay groundwork for meaningful digital transformation, and fuel later strategic decisions aimed at long-term sustainability.”
Local Media Association
Big Technology Podcast is the perfect guide to navigating what’s happening behind the scenes in Silicon Valley. In this episode journalist Alex Kantrowitz interviews Erik Brynjolfsson, the director of the Stanford Digital Economy Lab and professor at the Stanford Institute for Human-Centered AI. Brynjolfsson explains why fears that artificial intelligence would take human jobs haven't come to fruition…
A new category of artificial intelligence–“Foundational AI”–could have use in many industries from writing to coding to drug discovery. This episode of the weekly tech podcast from the Economist explores the foundation models and their implications for our future.
Check out the session archive from the BBC Academy’s June 2022 conference exploring the latest developments in the use of automation, algorithms and accountability in the news.
Food for Thought
Will this Fruit-picking Robot Transform Agriculture? Machines are now so advanced that they can patrol our streets, build cars, and sort goods in a warehouse. Having a machine able to do a simple task like picking an apple is unexpectedly hard: this article published by The Guardian is delving into this topic, explaining how these robots are complicated and at the same time a big benefit for farmworkers.
How Does AI Trained on 4chan Behave? Very Badly (as Expected) A YouTuber named Yannic Kilcher trained a bot using posts from the infamous 4chan Politically Incorrect board. The result is in line with the expectations: Kilcher trained the bot - GPT-4chan based on the GPT language models developed by OpenAI - with racist, misogynistic, and antisemitic messages. Then he released GPT-4chan bots that posted thousands of times on the 4chan /pol/ imageboard. “The model was good, in a terrible sense,” says Kilcher in a YouTube video describing the experiment.
Your Favorite Writers and Artists Are AI Programs A Vanity Fair feature article delves into the future of writing and artistic production. The thesis is pretty simple: a new generation of AI programs (Sudowrite, Dall-E, ShortlyAI for example) can create well-written articles and stunning illustrations. And this is just the beginning, in the next few years we will see AI pervading even more of the creative world.
Why AI is everywhere except your company It’s true. AI is changing our life and in the last few years the use of algorithms and artificial intelligence in conversations with friends and work colleagues is skyrocketing. However, according to an article published by the Financial Times, the uptake of artificial intelligence is slow and the adoption of AI is largely absent from most of the organizations with which we directly interact or work.
The initiative is funded by Knight Foundation.
Newsletter produced by Angelo Paura