Our weekly Briefing is free, but you should upgrade to access all of our reporting, resources, and a monthly workshop.
This week on Indicator
Craig was in Kuala Lumpur for the Global Investigative Journalism Conference. After attending a ton of sessions, he put together 29 OSINT tools and tips he learned, including a cool AI tool for working with satellite imagery.
Craig also dug into X’s new “About this account” feature and outlined how digital investigators can use the information in their work.
Our next live workshop for members is today at 1 pm ET. We’ll dig into how to find interesting and potentially confidential documents sitting on a publicly-accessible server or website. If you’re a member and want to attend, email us for the link. We’ll also share the video, transcript, and AI-generated notes on our Workshops page.
Deception in the News
The Verge used Google’s AI to generate an image of a second shooter at the scene of the assassination of President John F. Kennedy Jr.
📍 The Club de Periodistas in Mexico is a nonprofit that gives awards to journalists and says it has “promoted journalistic excellence since 1952.” But an investigation by Factchequeado and Animal Politico reveals that it “operates as a front for Russian disinformation and Kremlin-aligned narratives,” extensively republishing RT and Sputnik. Also, it gave an award to Alex Jones.
📍 The Verge found “few filters or guardrails” when using Google’s latest image generator to create photorealistic disinformation. Prompts like “show the White House on fire with emergency crews responding” required no jailbreaking, and also worked when I tested it on Wednesday. This laissez-faire approach would have been unimaginable in the company’s earlier bold-and-responsible days. — Alexios
📍 Lawmakers around the world keep toying with legislation to combat deceptive AI. In Australia, independent senator David Pocock is proposing that creators of nonconsensual deepfakes (not just those that are sexual in nature) be fined or sued. In the United States, a bipartisan proposal seeks to increase fines for AI-enabled fraud and criminalize deepfakes of federal officials. (The second feels dead on arrival, based on court decisions to date…). — Alexios
📍 Dutch public broadcaster NOS said that it would quit X, explaining that “the amount of hateful responses and disinformation on X is large and unrestricted. Also under our own messages, making us unintentionally help spread them."
📍 Singapore is demanding that Apple and Google do more to prevent scammers from spoofing government domains in text messages.
📍 Search results in smaller languages are often worse than those in English. Case in point: pro-Kremlin propaganda was among the top results when Fact Check Greece’s Thanos Sitistas searched for keywords related to the false claim that Ukraine’s president owns a villa in Wyoming. In comparison, the same search in English featured multiple fact checks at the top.
📍 Just in time for American Thanksgiving, Bloomberg spoke to food bloggers and found that “AI-generated ‘recipe slop’ is distorting nearly every way people find cooking advice online, causing consumers to waste time and money on unappetizing or impossible dishes.”
Tools & Tips
Our friends at My OSINT Training shipped a whole bunch of new and updated bookmarklets.
MOT’s free bookmarklets are indispensable for investigating online profiles. If you’re unfamiliar with bookmarklets and how to use them, I laid it out in detail here.
You can read all about the new/updated bookmarklets on the MOT site, but here’s the highlights:
They integrated X’s “About this account” information into their X bookmarklet.
They launched a new interface for bookmarklets and added a button to instantly copy all of the data.
They launched new bookmarklets for Pinterest, Poshmark, Snapchat, and Soundcloud.
There’s a new TikTok Date Revealer Bookmarklet, too.
📍 A recent Briefing looked at AI tools that can assist with geolocating an image. Here’s another one: Rhino Map. Register for free and you get some search tokens. (via Cyber Detective)
📍 Nico Dekens, aka Dutch OSINT Guy, wrote, “When HUMINT Forgot the Internet: Why Human Intelligence Needs OSINT More Than Ever.”
📍 The Wonder Tools newsletter recently shared five tips for using AI from Alexandra Samuel, an expert in the field.
Events & Learning
📍 There’s a free Dec. 4 launch event for Michelle Amazeen’s new book, Content Confusion: News Media, Native Advertising, and Policy in an Era of Disinformation. It’s in Boston and starts at 12 pm.
Reports & Research
📍 17% of Australians say they have seen an AI label applied to misinformation on social media, more than double the share from last year, according to a survey conducted by the country’s communications authority.
📍 Fascinating finding in a doctoral dissertation by Manu Singh Burson (h/t Madeleine Daepp): Politicians who appeared to have had a higher share of bot followers saw a steep decrease in news coverage after X’s 2022 API changes reduced their likely inauthentic reach.
📍 Ekō, a US-based nonprofit, conducted an investigation into OpenAI’s Sora 2 product and found that “the platform algorithm delivers teenage users inappropriate, racist, and violent content.”
One More Thing
Kingston upon Thames, a borough outside of London, recently installed two large, festive murals to celebrate the holiday season. Unfortunately, they’re filled with the stuff of AI-generated nightmares.
“The murals showed horrifyingly disfigured people, animals and even a demented snowman, harkening back to a not-so-distant past when AI image generators struggled to generate pictures of hands with five fingers,” wrote Victor Tangermann of Futurism.
Gaze upon its horrors. via https://bsky.app/profile/stephenoakman.bsky.social/post/3m5w4iqrkf22x
The murals were eventually torn down, but not for the reason you might expect.
Indicator is a reader-funded publication.
Please upgrade to access to all of our content, including our how-to guides and Academic Library, and to our live monthly workshops.
