🛡️ Truth in the Digital Age: A Quick Guide to Factchecking (By Rescue Ranger)

Status
Not open for further replies.

RescueRanger

Meme Lord
Joined
Sep 20, 2008
Messages
33,386
Reaction score
77,522
Reputation
35,359.3
Country of Origin
Country of Residence
Dear members! With so many members relying on social media as the primary source of their news, it’s getting harder to tell what’s real and what’s just noise. Whether it’s a viral "breaking news" tweet or a suspicious looking video, being digitally savvy is more important than ever.

Here is a breakdown of how to sharpen your "BS detector" and keep your posts moderator friendly and prevent the spread of lies and disinformation.
1773317737278.png
A. What are facts? Facts are the truthful answers to a reporter’s 5 key questions: who, what, when, where, and how. Facts may include names, numbers, dates, definitions, quotes, locations, research findings, historical events, statistics, survey and poll data, titles and authors, pronouns, financial data, institution names and spellings, and historical or biographical details attributed to anyone or anything. Facts are checkable.

B. What is factchecking?
Factchecking is the process of confirming the factual accuracy of certain statements or claims, in order to create and share accurate, evidence-based media that relies on high-quality, reliable primary and secondary sources.

C. What kinds of facts do people often get wrong? The most frequent mistakes occur in the spellings of names and institutions, and the attribution or wording of quotes. These errors can be relatively harmless — for example, a throwaway remark about Ben Franklin. Or they can be devastating — for example, listing the wrong person in a breaking news article about a bombing.

While the majority of factual errors are probably not nefarious, there are instances in which people may deliberately hide important facts or introduce inaccuracy. For journalists in these situations, three maxims are useful in finding the facts: ‘follow the money’, ‘consider the source’, and ‘who benefits?’ Remember, a reporter’s job is to find and share the facts that matter, even if people don’t like it.

D. Facts are only as good as their sources. There are two main types of sources: primary sources and secondary sources. Primary sources may include people, transcripts, videos, visitor logs, raw data, peer-reviewed scientific studies, recorded interviews, your own original research, and in-person observation. Secondary sources may include newspaper articles, magazine articles, and books. (Important note: unlike magazines, many books are not factchecked! If you’re using a book as a source, look for a bibliography or notes to track down an author’s sources, and then re-report if needed.)

As with all sources, watch out for inaccuracy, outdated information, and unconscious bias (for example, avoid disproven studies, or articles that talk about people ‘looting’ vs ‘finding’ and ‘rioting’ vs ‘protesting’). Avoid spreading inaccuracy, outdated information, or unconscious bias. Instead, try to increase the world’s supply of truth by shining a light on facts that matter.

1773317772416.png

1. Understanding the Enemy: Mis- vs. Disinformation​

Not all "fake news" is created equal. Understanding the intent helps you react appropriately:
  • Misinformation: False information shared without harmful intent (like your aunt sharing a fake health tip she genuinely believes is helpful).
  • Disinformation: False information created and shared intentionally to deceive, cause harm, or manipulate public opinion. See video below:
To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.


2. The Verification Checklist​

Before you hit "Share," run through these steps:
  • Verify the Source: * Credibility:Is it a known news outlet? Check the "About Us" page. Look for a history of corrections.
    • Bias: Every source has a lean, but is it extreme? Use tools like Ad Fontes Media or AllSides to see where a source sits on the political spectrum.
  • Cross-Reference: Never rely on a single source. If a major event happened, multiple reputable outlets (AP, Reuters, BBC) will be reporting on it. If only one obscure blog has the "scoop," be skeptical.
  • Analyze the Claim: Is the headline designed to make you angry or scared? Clickbait often uses "loaded language" to bypass your logic and trigger your emotions.

3. Looking for Evidence & Context​

A claim without a source is just a rumor.
  • Check the Links: Does the article actually link to the studies or documents it mentions?
1773317238597.png
  • Context Matters: Sometimes a quote is real, but it’s been stripped of the sentences around it to change its meaning. Always look for the full transcript or video.

4. Spotting "Synthetic" Media (AI & Deepfakes)​

AI-generated content is getting scarily good, but it still leaves "digital fingerprints." Here’s how to spot them:
  • For Images: * The Details:Look at hands (too many fingers?), teeth (too many or blurry?), and jewelry (earrings that don't match).
    • Backgrounds: AI often struggles with text in the background or architectural logic (stairs leading to nowhere).
  • For Videos (Deepfakes): * Unnatural Blinking:Does the person blink normally?
    • Skin Texture: Does the skin look too smooth or "plastic"?
    • Audio Mismatch: Watch the mouth closely—do the lip movements perfectly match the sounds?

5. How to Stop the Spread​

If you find a post that is false:
  1. Don’t Engage: Commenting "this is fake" can actually boost the post in the algorithm.
  2. Report It: Use the platform’s reporting tools for "False Information."
  3. Share the Truth: Instead of sharing the fake post, share a link to a verified fact-checking site (like Snopes or PolitiFact) that debunks it.

Think Before You Link

1. Always apply the "C.R.A.P." TEST​

Before sharing, ask yourself:
  • Currency: When was this posted? Old news is often shared as "new" to stir up trouble.
  • Reliability: Does the site have an "About" page? Do they cite their sources?
  • Authority: Is the author an expert or a random account with a bunch of numbers in their name?
  • Purpose: Is this meant to inform, or is it trying to sell me something or make me angry?

2. THE 3-SOURCE RULE​

Never trust a "bombshell" that only appears on one website.
  • Search the headline: If it’s true, at least three major, independent news organizations (e.g., AP, Reuters, BBC, Wall Street Journal) should be reporting it.
  • Check Fact-Checkers: Search the claim on Snopes.com, PolitiFact.com, or FactCheck.org.

3. SPOTTING AI & SYNTHETIC MEDIA​

AI makes mistakes. Look for these "glitches":
  • The Hands & Ears: Look for missing fingers, merged hands, or earrings that don't match.
  • The Eyes: Does the person have a "dead" stare? Do they blink naturally?
  • The Background: Look for "melting" objects, nonsensical text, or blurry edges where a person's hair meets the background.
  • The Audio: Listen for "robotic" pauses or a metallic tingle in the voice.
 
Last edited:
How to use Gemini to check for synthetic images:
1. Open Gemini and go to the App tab
1773318023769.png

2. Scroll to the bottom of the Gemini apps page and make sure that SynthID is enabled
1773318084737.png
3. Now go to Gemini and run a test on an AI image. For this I have prepared a test image in advance:
Gemini_Generated_Image_874oyk874oyk874o.png

4. We upload the image on Gemini Chat and use the following prompt. 1773318678712.png
5. This prompt is telling Gemini that we wish for it to verify if this image is made with AI and asking it to use SynthID to test for watermarks on the image.

6. Here is our result:
1773318768843.png

Notice: This check will only work for AI generated content that has a watermark on the image or video - most creators will now suppress or use methods to conceal the watermark to avoid detection. For this we can use other tools that I will mention further down below.
 
Tools and resources:

Please find a list of useful tools and resources you can use to verify news and content online before sharing:

Fact Checking:
  • 20-Minutes The 20-minute editorial team provides you with its “fake off” fact-checking section to combat fake news.
  • BBC Verify The BBC's in house fact-checking team who investigate video and clips and analysing data.
  • Boomlive Leading fact-checkers investigates news, public data, algorithms, social media posts, public statements of consequence.
  • Fact Check Website that aims to reduce the level of deception and confusion in U.S. politics.
  • Fact Check Explorer Gonzo's Python version of his fact check explorer tool.
  • FEAT Gonzo's FactCheck Explore Analysis Tool.
  • Google FCE Search facts about a topic or person.
  • Google Trends Google Trends is a website by Google that analyzes the popularity of top search queries in Google Search
  • HoaxBuster Collaborative platform against disinformation: its website allows you to identify hoaxes on the Web, that is to say hoaxes.
  • Internet Archive This tool aims to combat disinformation by making more first-hand, accurate information available to the public in a "digital library".
  • Open Measures Analyze hate and disinformation online.
  • Politifact PolitiFact.com is a fact-checking website that rates the accuracy of claims made by elected officials, politicians, and other public figures.
  • RAND Tools that fight disinformation online.
  • Snoopes Internet reference source for researching urban legends, folklore, myths, rumors, and misinformation.
  • Verfication Toolbox A toolbox designed to help simplify and streamline verification for beginner.
Image Verification:
  • AI or Not Detect AI-generated images & audio for your business - and yourself.
  • Content At Scale AI image detector will determine if any of them have been generated by AI.
  • EKagent Earthkit Agent is a multi-modal agent combines real-time information with state-of-the-art ML models to perform accurate geolocation and verification.
  • FID Fake Image Detector, a powerful tool for detecting manipulated image.
  • Forensically The magnifier allows you to see small hidden details in an image.
  • FotoForensics Forensics analysis tools to determine whether a picture was modified or manipulated.
  • GoogleGet helpful context with about this images you find online.
  • Hugging Faces This app is a proof-of-concept demonstration of using a ViT model to predict whether an artistic image was generated using AI.
  • Illuminarty Detect AI generated images, synthetic, tampered images and Deepfake.
  • InVid The InVID project develops a video verification platform to detect emerging stories and assess the reliability of video content spread via social media.
  • V7 Labs V7 has released a fake profile picture detector to help you identify if the person is real or not.
Further training resources:
  • Bellingcat A beginner's guide to social media verification.
  • First Draft Online training and resources.
  • Poynter Some free training in relation to Fact-Chacking.
  • Verification Handbook A definitive guide to verifying digital content for emergency coverage.
 
Status
Not open for further replies.

Users who are viewing this thread

Latest Posts

Back
Top