One red chat bubble with horns, and one white chat bubble with a halo.

The Washington Post

0118-helpdesk-gptuseful

STOCK IMAGE

For The Washington Post

help-badreview-topper

BRITAIN-ROYALS-KATE-HEALTH

Contributor

0225-helpdesk-AVNAI

An illustration of a parent talking to a child about a robot.

1114-helpdesk-aikids

An illustrated collage with candidates from the 2024 election and the capitol building.

helpdesk-aipledge

An illustration of a mechanic being squished by an apple.

1019-helpdesk-FIXPHONES

An illustration of a hand with an emergency sign.

1009-helpdesk-MISINFOGUIDE

An image of an open coloring book showing two pages. One one page is a woman's face, but her hand disappears into her neck. On the other, a woman grips a flower, but her fingers are twice as long as usual.

09xx-help-shoppingai

A collage with an eye, keyboard, and a chat bubble.

helpdesk-aieatingdisorders

An illustration of a laptop hugging the earth

0420-helpdesk-earthday

0322-helpdesk-AICONSPIRATOR

A robot replaces King Kong on top of a skyscraper in a scene from the classic monster movie 'King Kong'. 

0221-techfriend-aifear

Hands on a keyboard with messages and reactions popping up.

0208-helpdesk-aiabuse

Tyre Nichols Family Holds Press Conference

1223-techfriend-tiktok-vid

1208-helpdesk-techfriendai

Respond to prefers-color-scheme dark in app web views

Follow section front redirects in Spectrum

Enable NLP source in PCS for load testing

Switch off the News Alerts Survey Widget when we have enough responses

Poll to determine if new LUF updates are available

Allows the winning A/B test variant to be turned on without Spectrum deployment

Show infinite scrolling sections in Assembler MultiView

Show option for NLP article summaries in Action Bar

Show NLP summaries collection recirc on article pages

A general feature flag for Recipe Features that aren't yet released

Feature flag for media team features that aren't yet released

Use new Orca video player for preroll plugin in Spectrum

Show the snow tracker (instead of the heat tracker) on Spectrum

Big Tech news and how to take control of your data and devices

See more coverage in your recommendations

Technology

Tips delivered to your inbox twice weekly

The Tech Friend newsletter

Business and Tech alerts

Anyone with an internet connection can watch breaking news unfold in real time, or at least some version of it. Across social media, posts can fly up faster than most fact-checkers and moderators can handle, and they are often an unpredictable mix of true, fake, out of context and straight propaganda.

The risk is even higher during a war such as the one <a href="https://www.washingtonpost.com/world/2023/10/09/israel-hamas-war-gaza-violence/" target=_blank>unfolding in Gaza</a> or during an election.

The rapid spread of easily accessible <a href="https://www.washingtonpost.com/technology/2023/03/30/midjourney-ai-image-generation-rules/" target=_blank>AI tools</a> is mudding the waters even further. Look no further than the mystery of the <a href="https://www.washingtonpost.com/lifestyle/2023/03/27/pope-francis-coat-puffy-white-ai-fake/" target=_blank>pope in an expensive-looking</a> puffy coat or a fake social media post that was quickly debunked claiming there was an explosion near the Pentagon.

How do you know what to trust, what not to share and what to flag to tech companies? Here are some basic tools everyone should use when consuming breaking news online.

1. Know why something might be misinformation

Think about who would benefit from spreading confusing information during a news event, and brush up on specific narratives going around. During elections, for example, experts say to look out for conflicting information and conspiracy theories, baseless accusations and unfounded concerns about voter fraud that may benefit one political party or candidate. Propaganda is also used as a tool during armed conflict and can come through official news releases or feed through unofficial channels.

If something feels too outrageous or satisfying, regardless of whether it lines up with your views, pause and do more research. Misinformation is often accidentally spread by people who want it to be true. The people behind it could be trying to rile up supporters or create more tension between opposing sides of an issue, or generally destabilize a population. You may not always be able to find an accurate version of events, especially in fast-moving situations. In those cases, wait as long as it takes for additional information before sharing or drawing conclusions.

Not all misinformation is serious. Some is created just for fun or to troll people, so be as skeptical of silly stories as serious ones.

Do not hit that share button. Social media is built for things to go viral, for users to quickly share before they are even done reading the words they are amplifying. No matter how devastating, enlightening or enraging a TikTok, social media post or YouTube video is, you must wait before passing it on to your network. Assume everything is suspect until you confirm its authenticity.

3. Check the source; don’t always trust “verified” accounts

Look at who is sharing the information. If it’s from friends or family members, don’t trust the posts unless the information is from someone who is there on the ground or from a confirmed expert. If it’s a stranger or an organization, remember that a verified check mark or being well known does not make an account trustworthy. Plenty of political pundits and big-name internet characters are posting inaccurate information right now, and it’s on you to approach each post with skepticism.

If the account posting something is not the source of the words or images, investigate where it came: dig to find the original Facebook, YouTube or social media account that first shared it. If you can’t determine the origin of something, that’s a red flag. Be wary of screenshots, which can be even harder to trace back, or anything that elicits an especially strong emotional reaction. Disinformation preys on that type of response to spread.

When screening individual accounts, look at the date it was created, which should be listed in the profile. Be wary of anything extremely new (say, it was created in the past few months) or with only a few followers. For a website, you can see what year it was created on Google. Search for the name of the site, then click on the three vertical dots next to the URL in the results to see the date it was first indexed by the search engine. Again, avoid anything too new. And don’t skip the basics: Do a Google search for the person or organization’s name.

Doing mini background checks on every random social media account is extremely time-consuming, especially with new content coming from so many places simultaneously. Instead, trust the professionals. Legitimate mainstream news organizations are built to vet these things for you, and often do report on the same videos or photos taken by real people after they have confirmed their origin.

Use a dedicated news tool such as Apple News, Google News or Yahoo News, which choose established sources and have some built-in moderation. On social media, make or find lists of vetted experts and outlets to follow, specifically for news about the topic you’re following.

If you consume breaking news on X, formerly known as Twitter, take care to follow confirmed reporters from trusted outlets on the ground. <a href="https://www.washingtonpost.com/technology/2022/11/01/twitter-verification-blue-check-mark/" target=_blank>Changes to Twitter’s</a> verification system have made this more difficult. Do not assume that an account is trustworthy or vetted by X because it has a blue check mark or because it was <a href="https://www.washingtonpost.com/technology/2023/10/08/israel-hamas-disinfo-musk-twitter-x/" target=_blank>shared by the company’s owner</a>, Elon Musk. Visit news sites directly or look up the people you trust on another platform.

5. Seek out additional context about news events

Misinformation often spikes before, during and after major news. Real-time news events will include a flow of disjointed information from the ground, such as smartphone videos and first-person narratives. Even if you see only legitimate posts, they can still be confusing or misleading — they may be the most compelling pieces of a puzzle, but they are not the whole picture. Try to augment any one-off clips or stories with broader context about what is happening. Mix in information from established experts on the topic, whether it’s foreign policy, cyberwarfare, history or politics. You can also turn to online or television outlets that add this context.

AI image generation tools make identifying fakes much harder, but not impossible. Here are five clues that you can look for to spot AI-generated images, including by zeroing in on hands, background images and inanimate objects that often don’t look quite right. You can find even more tips in our <a href="https://www.washingtonpost.com/technology/2023/03/31/tips-spot-ai-fake-images/" target=_blank>Tech Friend newsletter</a>.

Look at the hands. AI software has a history of generating human hands with too many fingers or other oddities, though the technology is getting better fast.

Zoom in on any inanimate objects in the image to see whether they feel off. Focus on items in an image like eyeglasses, fences or bicycles and see whether they have any telltale flaws.

If you’re wondering whether an image is made by AI, look for writing on such objects as street signs or billboards and see whether it’s backward or nonsensical.

Scan the background. AI-generated images may have blurry or distorted details, particularly in the background.

Check whether the images are overly glossy or artistic-looking. Some AI-generated images of real people appear garishly stylized or depict people with plastic-looking faces.

Use all the other misinformation signs if you’re still unsure. Is it shared by a reliable source? Are there separate reports of the same event?

If you’re interested in doing deeper dives into unverified reports, start with <a href="https://www.washingtonpost.com/politics/2021/03/19/how-spot-fake-video/" target=_blank>this extensive guide </a>on how to screen videos. Look for multiple edits and odd cuts, or listen closely to the audio and run it through a third-party tool such as InVid, which helps check the authenticity of videos. This can be harder on live-streamed videos, like what’s on Twitch or other live social media option.

To check images that aren’t AI-generated but still seem suspicious, put them into Google’s image search by grabbing a screenshot and dragging it to the search field. If it’s an old image that has circulated before, you may see telling results.

Social media sites do have some of their own fact-checking tools and warning labels, and many have added <a href="https://www.washingtonpost.com/technology/2020/10/20/election-social-misinfo/" target=_blank>special sections</a> to promote official election results. However, given the sheer volume of posts with which they are dealing, a problematic video or post can still be seen by millions before ever getting flagged.

Keep an eye out for content warnings on social media sites for individual posts, which can appear as labels below links or as warnings before you post something that could be misleading. Look up individual stories or images on fact-checking sites such as <a href="https://www.washingtonpost.com/news/fact-checker/" target=_blank>The Washington Post’s Fact Checker</a>, <a href="https://www.snopes.com/" target=_blank>Snopes</a> and <a href="https://www.politifact.com/" target=_blank>PolitiFact</a>.

Shira Ovide and Geoffrey A. Fowler contributed to this report.

New York University, BA in journalism

Heather Kelly is a San Francisco-based reporter covering the ways technology affects everyday life. 

Heather Kelly

washpost

How to avoid misinformation, fake AI images on X, TikTok and social media

the-washington-post

Illustration by Elena Lacey/The Washington Post

Tech in Your Life

Future of Work

Your Data and Privacy

Internet Access

What's New

Ethical Issues

Ask a Question

Help Desk

prism://prism.query/site,/technology/consumer-tech/ethical-issues&limit=20

prism://prism.query/site,/technology/consumer-tech&limit=20

prism://prism.query/two-sites,/technology,/politics/powerpost/the-technology-202?limit=20

How to avoid falling for misinformation, AI images on social media

Help Desk: Making tech work for you