Journalists can now identify Spanish-language deepfake audio using a new tool called VerificAudio launched Tuesday, according to an announcement from Prisa Media.
“In a new disinformation ecosystem where cloned voices are now almost imperceptible and in 2024 50% of the population will vote, today we are launching an application project from #IA to #audio: VerificAudio, a tool to detect Audio deepfakes,” Ana Ormaechea, Prisa Radio’s chief digital officer tweeted.
En un nuevo ecosistema de desinformación donde las voces clonadas son ya casi imperceptibles y en 2024 vota el 50% de la población, lanzamos hoy un proyecto de aplicación de la #IA al #audio: VerificAudio, una herramienta para detectar deepfakes de Audio: https://t.co/rLw00Xcz80
— ana ormaechea (@aormaechea) March 12, 2024
VerificAudio uses both identification and comparison models, which means that when journalists upload an audio file, the tool can identify the probability of the audio being real or fake. Then, if the user has a real audio clip of the voice they’re testing, the comparison system can determine if the file being tested is cloned or not.
Prisa Media, which owns El País, El HuffPost, financial newspaper Cinco Días, sports daily AS, and 1,200 radio stations across Spain and Latin America, developed the tool with the Google News Initiative and Spanish technology company Minsait. The tool is currently available to Prisa’s newsrooms in Mexico, Chile, Colombia, and Spain, and is working on rolling it out to all of its journalists around the world.
“This project that was born in [Colombian media company] Caracol and is progressively spreading to all our newsrooms, is part of PRISA’s commitment to the fight against disinformation and advocates for quality journalism,” José Gutiérrez, Prisa Media’s director of solutions, digital and technology, said in the release. “Technology has become a weapon to fight disinformation and deepfakes and is also a new essential tool for journalists. We believe that with the development of artificial intelligence many opportunities are created, but also many dangers that must be faced as they are detected.”
On VerificAudio’s website, users can test whether or not they can distinguish a real voice from a faked one. Users can also upload audio files they want checked for Prisa Media journalists to investigate (WhatsApp audio files are downloadable, FYI!).
Find the VerificAudio launch information here and its methodology here.
Google is one of the most important vectors for online spammers and scammers. The search engine’s dominance means that, if you want to generate traffic — a.k.a. potential marks — you’ll probably need to convince Google’s algorithm to send it your way. And there’s a good chance the resulting traffic will be monetized using Google’s own ad tools, which has a knock-on effect on the rest of the digital ad market. So the tech giant’s responses to new iterations of bad behavior have a big impact on the broader web publishing world.
On Tuesday, Google announced an important set of algorithm changes in an attempt to deal with a wave of AI-generated spam:
In 2022, we began tuning our ranking systems to reduce unhelpful, unoriginal content on Search and keep it at very low levels. We’re bringing what we learned from that work into the March 2024 core update.This update involves refining some of our core ranking systems to help us better understand if webpages are unhelpful, have a poor user experience or feel like they were created for search engines instead of people. This could include sites created primarily to match very specific search queries.
We believe these updates will reduce the amount of low-quality content on Search and send more traffic to helpful and high-quality sites. Based on our evaluations, we expect that the combination of this update and our previous efforts will collectively reduce low-quality, unoriginal content in search results by 40%.
(Here’s coverage from Wired, The Verge, Ars Technica, Gizmodo, and TechCrunch.)
The update has three prongs, which Google is calling expired domain abuse, scaled content abuse, and site reputation abuse.
Expired domain abuse involves someone buying a domain that Google had previously considered high quality and coasting on that reputation to drive traffic to, er, low-quality material. This is not a new problem; longtime readers might remember, in 2013, when a spammer grabbed the domain name of the Online Journalism Review and filled it with links to their products. Expired news sites are a particularly juicy target; New York Times reporter Lydia DePillis recently mourned The Washington Independent, an early-Obama-era politics site that has been turned into, well, something else.Site reputation abuse is similar, except rather than taking an entire domain, the spammer manages to post their own low-quality content onto an existing high-quality site — usually by exploiting a security hole in their content management systems or taking advantage of some long-forgotten user-generated-content section. Those who control these sites have two months to clear our the bad stuff before Google brings the hammer fully down.
And scaled content abuse is mostly driven by AI. A scammer uses something like ChatGPT to generate oceans of mealy-mouthed content relating to some topic or another in hopes of drawing organic search traffic. Sometimes they’ll pull an existing site’s sitemap to specifically copy its content (in just-altered-enough form) and steal away its traffic. (Some are even dumb enough to brag about their heist online!) The new policy “will allow us to take action on more types of content with little to no value created at scale, like pages that pretend to have answers to popular searches but fail to deliver helpful content,” Google says.
These changes sound on net like good news for legit publishers — any traffic redirected from spam sites is traffic that might instead be directed their way. But while you’re waiting for that organic boost: Renew your domain names, people.
But Report for America’s approach to those newsrooms is changing. At a Journal-isms roundtable last week on “Surviving the collapse of traditional business models,” Kim Kleman, Report for America’s executive director, told attendees that while the organization has had partnerships with hedge fund-owned news organizations in its time operating, “we are phasing those out.” Journal-isms, a blog about media and diversity, confirmed the news in an article published this weekend.
Kleman’s comments came in response to a question from Tracie Powell, founder and CEO of the Georgia-based Pivot Fund, which maps, funds, and supports news organizations trusted by people of color. Powell emphasized that while the Pivot Fund supports both for-profit and nonprofit outlets, like Report for America, Pivot specifically focuses on investing in “independently owned for-profit organizations — community-based, and serving, organizations.” She then asked Kleman, “Do you think there is an issue with nonprofit dollars going to corporate news organizations, specifically those that are owned by hedge funds?”
Kleman responded, “We started at Report for America working with what are now hedge fund-owned organizations because we considered ourselves as supporting communities, and sometimes, those newsrooms owned by those organizations, those companies, were the only game in town.”
She continued, “We have seen so much need over the years by independent for-profit [and] certainly nonprofit organizations that we have some newsrooms that are owned by hedge funds, but we are phasing those out.” (A clearly surprised Richard Prince, a longtime columnist writing about diversity issues in media and the roundtable host and moderator, interjected, “Oh really!” You can listen to the comments yourself in the recording around the 50-minute mark.)
Steven Waldman, Report for America co-founder and former president, and fellow co-founder Charles Sennott had both previously defended their decision to place reporters in such newsrooms. In 2020, for instance, Waldman told Dan Kennedy, “we have on occasion accepted applications from newspapers with the problems you mentioned if we were convinced that they would use the reporter to better serve their readers. If we can be a positive force in helping those newspapers tip more in the direction of great journalism, we view that as a real positive step.”
But in a way, this full break with hedge fund-owned outlets seems like the natural conclusion of a shift away from hedge fund-owned publications that RFA began in 2021. Despite defending RFA’s rationale for funding reporters in these newsrooms, Waldman also made no secret of his qualms about the damage hedge funds can do to local news. In late 2021, one of those hedge fund-owned chains took offense. Columbia Journalism Review reported then that Waldman’s LA Times op-ed questioning the viability of hedge fund ownership as a business model for sustainable, robust local news led McClatchy (controlled by the hedge fund Chatham Asset Management since 2020) to choose not to renew its partnership with RFA. At the time, 31 RFA journalists were reporting across 21 McClatchy newsrooms, per CJR.Some local and national journalists, among others, voiced praise and critiques of Report for America’s decision on Twitter (or X).
I empathize with folks saying “Many local newsrooms will suffer.”
But maybe RFA has an obligation not to send (mostly) young and naive reporters to hedge fund newsrooms where they are more likely to be exploited and discarded. See: me, McClatchy, 2021. https://t.co/luWR5MyTm7
— Haley Samsel 🍂 (@haley_samsel) March 4, 2024
This would be totally destructive in Wisconsin. @Report4America, I emailed you. I am not a major donor, but, please, recant!
Many favorite journalists are or were.@danielle_duclos @madeline_heim @caitlooby @_CleoKrejci @MadisonLammert0 @vaisvilas_frank @ReneeNHickman @SarahVolp https://t.co/L8RCLfoiBc
— Quinton Klabon (@GhaleonQ) March 4, 2024
I feel for the folks in these newsrooms, many of whom are excellent, exhausted journalists trying to plug the multiplying holes in their coverage as best they can. But given Alden’s track record in particular, it’s hard for nonprofits to justify sending resources to those papers https://t.co/V435GAEYK9
— McKay Coppins (@mckaycoppins) March 4, 2024
Terrible idea. This hurts readers and journalism. Not Gannett, Alden, etc. https://t.co/L0wmTqHRKF
— Corri Hess (@CorriHess) March 4, 2024
I’m so happy to announce that we’ve added two new full-time writers to the Nieman Lab team. Both of them are starting today.
Andrew Deck is our new AI staff writer, focused on the intersection of generative AI and journalism. Previously, he was a staff reporter at Rest of World, an international nonprofit publication he helped launch in 2020 focused on the impact of technology in non-Western countries. Andrew began his career in Japan, where he contributed to English-language publications including The Japan Times and was an assistant editor at the Tokyo-based magazine Metropolis.
Andrew is interested in covering labor displacement in newsrooms, training data contracts between media companies and AI developers, issues reporting on synthetic media, and overall the ethics of using AI in the editorial process.
Neel Dhanesha is our new general staff writer. He is a writer and audio producer from Bangalore. Before joining the Lab, he was a founding staff writer at the climate news startup Heatmap. He has also been a reporter at Vox, a fellow at Audubon magazine, and an assistant producer at Radiolab, where he helped produce The Other Latif, a series about one detainee’s journey to Guantanamo Bay. He is a graduate of the Literary Reportage program at NYU.
Neel intends to rove widely in his coverage for the Lab, but is especially interested in writing about science and climate reporting and the audio industry.
Andrew and Neel both live in Brooklyn and will be working for Nieman Lab remotely full-time. (This means we can now say we have an NYC bureau, IMO.) They join Sarah Scire, Josh Benton, Sophie Culpepper, Hanaa’ Tameez, and me. We’re so excited to have them on board.
You know a publication that’s never had that problem? The French satirical newspaper La Bougie du Sapeur, founded in 1980. That’s because it only publishes an issue once every four years, on February 29. That being today’s date, Issue 12 has just hit Parisian newsstands.
Je l'ai ! #29Fevrier
La Bougie du Sapeur,
L'unique « quotidien » à paraître tous les quatre ans. pic.twitter.com/74AT1Hehhc— Julien Cottereau (@j_cottereau) February 29, 2024
Jean d’Indy, le rédacteur en chef du journal satirique «La bougie du sapeur», périodique publié tous les 29 février, présente sa 12e édition. Pour le meilleur et pour le rire, l’intelligence artificielle fait la une. pic.twitter.com/gR1uwqt0xE
— Le Figaro (@Le_Figaro) February 29, 2024
La Bougie du Sapeur de retour, comme tous les quatre ans. pic.twitter.com/7YoQNjKuge
— LeBlogTVNews (@Leblogtvnews) February 29, 2024
I’d link to the newspaper’s website, but it doesn’t have one. (If it did, though, I’m sure it would have a much higher metabolism — maybe posting every 18 months or so.) When February 29 falls on a Sunday, La Bougie du Sapeur also publishes a special Sunday supplement; the first came out in 2004, with the next scheduled for 2032. Today’s top story: the rise of artificial intelligence. Can’t wait for Part 2!
Editor Jean d’Indy describes the content as “French humor, and it does not translate into other languages. We try to be silly but not nasty. To poke fun without being cruel.” Print runs are around 200,000, and the newsstand price is €4.90 — about $5.30 American. (Your best deal, though, is the 100-year subscription for €100.) And it’s all profitable.
American newspapers continue to cut print days — from the traditional 7 issues a week to 6, or 3, or even 1. Maybe La Bougie du Sapeur has figured out the end game.
The Nieman Journalism Lab is a collaborative attempt to figure out how quality journalism can survive and thrive in the Internet age.
It’s a project of the Nieman Foundation for Journalism at Harvard University.