Fergal Carr’s Post

Interesting survey of the various applications of LLMs in journalism, from thoughtless to thoughtful: "Faced with the chaotic, messy reality of everyday life, LLMs are useful tools for summarizing text, fetching information, understanding data, and creating structure. But always, always with human oversight: guiding the summaries and then checking the results, teaching the bot to navigate an audit report, deciding when and when not to put the bot in control, and designing the rest of the software that powers the site. In all of these cases, it's humans first and humans last, with a little bit of powerful, generative AI in the middle to make the difference." https://lnkd.in/g5vXs5vz

As a newsroom veteran now writing about lots of cutting edge tech, I have a lot of natural skepticism about the hype around the current wave of AI, especially as it pertains to journalism. So who better to write about how Virginia Tech is working with The Washington Post to deliver an AI-driven tool for users to be able to search their entire news archive, and how they’re navigating the pitfalls that have plagued so many of the current LLMs. https://lnkd.in/eiD8bJpN

In this overview, Cameron R. Wolfe, Ph.D. shows the basic components of AI-powered search systems, including lexical search, bi-encoders, and cross-encoders:

File based RAG (retrieval augmented generation) is all the rage, but I'm really interested in website RAG... so many use cases! Imagine being able to automatically ingest and analyze entire websites, like a local government site, to surface key insights and answer questions. 🤖📊 With website RAG, you could: - Instantly search through years of council meeting minutes - Analyze budget trends and spot anomalies - Generate summaries of complex policy documents ...and so much more! The possibilities are endless for empowering journalists, researchers, and engaged citizens to understand and interact with public data in new ways. 📰🔍 I've got the AI UI in place to manage it, now I've just got to get good at scraping and pre-processing the data... #AI #RAG #CivicTech #Journalism #OpenData

Join this upcoming webinar to learn how our AI-powered search and discovery capabilities efficiently classify different types of documents (news articles, reports, transcripts, fillings and more) so you never miss the market moving information that matter most to you. https://lnkd.in/gQqNMT-b

Join this upcoming webinar to learn how our AI-powered search and discovery capabilities efficiently classify different types of documents (news articles, reports, transcripts, fillings and more) so you never miss the market moving information that matter most to you. https://lnkd.in/dBvzemvk

Journalists are expressing concerns as more publications begin to sign content deals with AI firms. The primary issues revolve around the potential impact on their roles, job security, and the overall quality of journalism. These partnerships, while promising efficiency and cost savings, raise questions about the future of human journalists in an increasingly automated industry. Many are wary of how AI-generated content might influence the integrity and depth of news reporting. As this trend continues, it remains to be seen how the balance between technology and traditional journalism will evolve. #Journalism #AI #Media #JobSecurity #TechInMedia ---------------------- Learn more here: https://lnkd.in/e3QkBZFK

Join this upcoming webinar to learn how our AI-powered search and discovery capabilities efficiently classify different types of documents (news articles, reports, transcripts, fillings and more) so you never miss the market moving information that matter most to you. https://lnkd.in/escMBhs3

This is an eye-opening article in the WSJ on the dangers of generative AI being used to create partisan "news" sites. It's alarming to see how easy and inexpensive it has become for anyone to set up an AI-powered disinformation machine capable of churning out thousands of slanted articles per day without crediting the original sources.  In the article, the author created a fully functioning partisan AI news site designed to influence a real U.S. Senate race. The site was up and running in just two days for barely over $100, which is a wake-up call. Are we witnessing the Newspeak that Orwell warned us about? Or it could be a new asset class ... "Like rental properties and dividend stocks, AI content farms are an easy passive-income investment. The economics are straightforward. Setting up an AI content farm costs about $100, and ongoing costs include about $5 a month for a web-hosting company and the expense of running ChatGPT’s application-programming interface (API). That cost depends on the number of articles published; based on OpenAI’s listed API prices, my estimate is $3 or less a month to publish 50-100 articles a day." per the author.

🚀 Your AI toolkit just got a major upgrade! I updated the Journalists on Hugging Face community's collection with tools for investigative work, content creation, and data analysis. New additions include: - Jonathan Soma's excellent 6-part video series on AI for investigative journalism - Jeremy Caplan's curated AI Spaces on HF - Joshua Lochner's Whisper Timestamped (with diarization!) for private, on-device transcription - Flux models for image gen & LoRAs - Finegrain's object cutter and object eraser (this one's cool) - FineVideo: massive open-source annotated dataset + explorer Miquel Àngel Farré - Qwen2 chat demos, including 2.5 & multimodal versions (crushing it on handwriting recognition) - GOT-OCR integration - HTML to Markdown converter - Text-to-SQL query tool by David Berenstein for HF datasets There's a lot of potential here for journalism and beyond. Give these a try and let me know what you build! You can also add your favorite ones if you're part of the community or share them in the comments! 👉 Check it out: https://lnkd.in/evTJDQUG #AIforJournalism #OpenSourceAI