I’m continuing my series of round-ups on artificial intelligence and journalism. Here are some recent links you might want to read.
Reuters/Oxford and Journalism AI reports on what’s happening in newsrooms
The Reuters Institute and University of Oxford interviewed more than 40 new organizations from across the globe to uncover editorial strategies, deployment approaches, technical infrastructure needs, organizational needs and what might come next. Read the full report from The Reuters Institute.
JournalismAI spoke with hundreds of editors, journalists and technologists to inform their second global survey of what news organizations are doing with AI. The group is a collaboration between the journalism program at the London School of Economics and the Google News Initiative. Read the whole report on the JournalismAI website or the coverage of key findings on The Verge.
The Financial Times visually explains AI
We’ve all started using ChatGPT and other large language models, but how do they work? My favorite explanatory metaphor is to call them “super auto complete.” The Financial Times digs into the transformer models under the hood and visually explains how they are trained. Interact with the full story from the FT visual storytelling team.
More news sites blocking OpenAI scraping
Journalist Ben Walsh is tracking the robots.txt files of more than 1,000 news domains. This is the file websites use to indicate to web crawlers what they are allowed to access. He reports on X, formerly known as Twitter, that almost half are now including instructions to OpenAI that they should not scrape content from their websites.
Piloting AI tools for local newsrooms
With support from the Knight Foundation, the Associated Press wrote a report last year about AI in local news. Now, they are supporting five pilot projects that aim to put AI technologies in service of the needs of smaller local newsrooms. These include police activity summaries, automatic weather alerts, video transcriptions and summaries, city council meeting keyword extraction and editorial coverage planners. Read more on the NiemanLab blog.
Publishing cutting-edge research from academics
A new study from Journalism & Media looked at reader trust in AI-human reporting. Key findings relate to human control over integrations. Read the full paper open-access on the journal website.
If you’re an academic and do similar research on how journalists and newsrooms navigate the AI hype and practical needs, checkout the call-for-papers for the Digital Journalism special issue. Submit an abstract by Jan. 10. See more from Professor Tomás Dodds on X.
That’s all for this round-up. For more, check out our earlier update. Found something relevant and want to share? Add a comment below or drop me a line at @rahulbot@vis.social or @rahulbot@twitter.com.
Originally published at https://www.storybench.org on October 18, 2023.
