Deepak Adhikari - Media Helping Media https://mediahelpingmedia.org Free journalism and media strategy training resources Sun, 24 Dec 2023 09:48:53 +0000 en-GB hourly 1 https://wordpress.org/?v=6.4.4 https://mediahelpingmedia.org/wp-content/uploads/2022/01/cropped-MHM_Logo-32x32.jpeg Deepak Adhikari - Media Helping Media https://mediahelpingmedia.org 32 32 How to detect AI-generated images https://mediahelpingmedia.org/advanced/how-to-detect-ai-generated-images/ Sun, 24 Dec 2023 09:22:31 +0000 https://mediahelpingmedia.org/?p=2923 Fact-checking journalist Deepak Adhikari, the editor of Nepal Check, shares how he and his colleagues combat the spread of fake AI images on social media and in other news output.

The post How to detect AI-generated images first appeared on Media Helping Media.

]]>
Image of robot and smartphone by Matt Brown (https://www.flickr.com/photos/londonmatt/) released via Creative Commons BY DEED 2.0 (https://creativecommons.org/licenses/by/2.)
Image of robot and smartphone by Matt Brown released via Creative Commons BY DEED 2.0

Fact-checking journalist Deepak Adhikari, the editor of Nepal Check, has shared a piece he wrote about the spread of AI-generated image following an earthquake in Nepal in November 2023. The article, below, explains how his organisation and others set about identifying the fake photographs. Deepak hopes the methods he and his team used will be of use to other journalists trying to combat the spread of fake images on social media and in other news output.


Social media is flooded with AI-generated images. Here’s how to detect them

By Deepak Adhikari, editor of Nepal Check

Following the devastating earthquake that struck Jajarkot district in Karnali Province in early November, social media users shared AI-generated images claiming to show the devastation caused.

One photograph showed dozens of houses ruined by the earthquake with people and rescuers walking through the debris. The photo was initially shared by Meme Nepal. It was subsequently used by celebrities, politicians and humanitarian organisations keen to draw attention to the disaster in one of Nepal’s poorest regions.

The image was used by Anil Keshary Shah, a former CEO of Nabil Bank and Rabindra Mishra, a senior vice president of National Democratic Party. (See archived version here and here)

When Nepal Check contacted Meme Nepal in an attempt to find the original source of the photo, they replied that they had found the image on social media.

A screenshot of Arjun Parajuli’s post on Facebook along with a poem lamenting the scene from the image
A screenshot of Arjun Parajuli’s post on Facebook along with a poem lamenting the scene from the image

A month on, the AI-generated images supposedly showing the aftermath of the earthquake continued. On December 14, 2023, Arjun Parajuli, a Nepali poet and founder of Pathshala Nepal, posted a photo claiming to show students studying in the ruins of the earthquake at Jajarkot. Parajuli. The poet attached the photo to a poem, had reshared the image from Manish Khadka, who identifies himself as a journalist based in Musikot of Rukum district.

A screenshot of Manish Khadka’s post on Facebook with a caption claiming to show students in Rukum and Jajarkot
A screenshot of Manish Khadka’s post on Facebook with a caption claiming to show students in Rukum and Jajarkot

Both these viral and poignant images were fake. They were generated using text-to-image generator platforms such as Midjourney, DALLE.

In the digital age it’s easy to manipulate images. With the rise of AI-enabled platforms it’s possible to generate images online quickly and convincingly. AI-generated images have evolved from amusingly odd to realistic. This has created further challenges for fact-checkers who are already inundated with misleading or false information circulating on social media platforms.

Fact-checkers often rely on Google’s Reverse Image Search, a tried and tested tool used to detect an image’s veracity. But Google and other search engines only show photos that have been previously published online.

So, how can one ascertain if an image is AI-generated? Currently, there is no tool that can determine this with 100% accuracy.

A screengrab of result on ISITAI after uplaoding the viral image on the platform
A screengrab of result on ISITAI after uplaoding the viral image on the platform

For example, Nepal Check used Illuminarty.ai and isitai.com to check the earthquake images to try to find out if they were generated using AI tools. After uploading an image to the platforms a percentage of how likely the image is to be generated by AI is shown.

A screengrab of result on Illuminarty after uplaoding the viral image on the platform
A screengrab of result on Illuminarty after uplaoding the viral image on the platform

Nepal Check contacted Kalim Ahmed, a former fact-checker at AltNews. He made the following observations about the image claiming to show devastation of the earthquake in Jajarkot.

  • If you zoom in and take a closer look at the people they appear deformed and like toys.
    The rocks/debris just at the centre look like they’re straight out of a video game made in the late 90s or early 2000s.
  • In the absence of a foolproof way to determine whether a photo is AI-generated, using observational skills and finding visual clues is the best way to tackle them.
Examination of an AI image
Examination of an AI image

A healthy dose of skepticism about what you see online (seeing is no longer believing), a search for the source of the content, whether there’s any evidence attached to the claim, and looking for context are powerful ways to separate fact from fiction online.

Further examination of an AI image
Further examination of an AI image

In a webinar in August this year organised by News Literacy Project, Dan Evon urged users to keep asking questions (is it authentic?). With the AI-images, their surfaces seem unusually smooth, which can be a giveaway, according to him. “Everything looks a little off,” he said.

Dan suggests looking for visual clues, adding that it was crucial to find out the provenance of the image. Experts caution that the virality of content on social media often stems from its ability to generate outrage or controversy, highlighting the need for careful consideration when encountering emotionally charged material.

In her comprehensive guide on detecting AI-generated images, Tamoa Calzadilla, a fellow at the Reynolds Journalism Institute in the US, encourages users to pay attention to hashtags that may indicate the use of AI in generating the content.

While AI has made significant progress in generating realistic images, it still faces challenges in accurately replicating human organs, such as eyes and hands. “That’s why it’s important to examine them closely: Do they have five fingers? Are all the contours clear? If they’re holding an object, are they doing so in a normal way?”, Tamoa writes in the guide.

Experts recommend that news media disclose information to readers and viewers regarding AI-generated images. Social media users are also advised to share the process publicly to mitigate the spread of misinformation.

Although the images purporting to depict the earthquake in Jajarkot lack a close-up view of the subjects, upon closer examination it becomes evident that they resemble drawings rather than real humans. Nepal Check also conducted a comparison between the viral AI-generated images and those disseminated by news media. We couldn’t find any such images that had been published on mainstream media in the aftermath of the earthquake.

The post How to detect AI-generated images first appeared on Media Helping Media.

]]>
Creating a strong fact-checking system https://mediahelpingmedia.org/advanced/creating-a-strong-fact-checking-system/ Sat, 04 Feb 2023 06:47:56 +0000 https://mediahelpingmedia.org/?p=2636 It’s the job of the journalist to try to find and present the truth, but fact-checking isn't easy. It requires a methodological approach to verification. 

The post Creating a strong fact-checking system first appeared on Media Helping Media.

]]>
Fact-checking training Kenya. Image by Media Helping Media released via Creative Commons
Fact-checking training Kenya. Image by Media Helping Media released via Creative Commons BY-NC-SA 4.0

Many find it difficult to distinguish between credible and misleading content in today’s digital age. With the rise of social media, false and misleading information has become more prevalent, making it even more difficult for people to tell the difference between fact and fiction.

It’s the job of the journalist to try to find and present the truth, but fact-checking isn’t easy. It requires a methodological approach to verification.

Because many media outlets are profit-driven, sensationalist and clickbait content frequently takes precedence over fact-based journalism. As a result, trust in journalism as a whole has declined, while, at the same time, skepticism about the reliability of information published and shared has grown.

The preference of social media platforms for engagement and amplification has contributed to the information disorder. Indeed, credible and evidence-based information is in short supply, while false and misleading content continues to spread on social media.

Because of this it is critical to have a strong fact-checking system in place to combat the spread of false and misleading information.

A fact-checker must be skilled at digital investigation and willing to go the extra mile to uncover a kernel of truth. To assess the veracity of information accurately, fact-checkers must be trained in and have access to digital tools, techniques, and resources.

Fact-checking in action

At Nepal Check, we use both human intelligence and digital tools to verify information spreading online. Over the past six months, we have fact-checked a wide range of false claims, from political statements to health misinformation.

On the eve of the November 20 2022 elections in Nepal, we fact-checked screenshots purported to be from a secret circular issued by the ruling Nepali Congress urging its supporters not to vote for Maoist candidates.

Not only was the so-called “secret circular” made up. Election misinformation spreaders created screenshots of ‘news’ that claimed to have been published by reputable digital outlets and a newspaper in Nepal.

We dug deep into the misinformation spreaders to find out what party they were affiliated with. We found that the majority were affiliated with an opposition party that competed with the ruling alliance.

One of the fake news reports had a reporter’s byline. The claim was refuted by the reporter himself. In the fact-check, Nepal Check cited him. We also cited news outlets that claimed the screenshots were fake. In addition to relying on their statement, we checked the claims on news websites.

Finally, our investigation revealed that the Nepali Congress party’s Central Working Committee had not met in nearly three months when the meeting was claimed. We also used the occasion to explain how morphed screenshots are created. In doing so, we hoped to raise awareness about the technique so that people would think twice before spreading false information.

We fact-checked videos with false claims that Gagan Thapa, the general secretary of the Nepali Congress, promised freebies if elected prime minister. Thapa is a popular politician and is often the target of misinformation. The video claiming Thapa promised free smartphones, free petrol and a monthly stipend for the elderly had circulated on social media for several months.

To debunk the claim, we needed to find the original, full version of the video. So we started by looking up the video on YouTube. We discovered a video of the political function, but it lacked the section where Thapa was quoted. However, the video provided an important clue that a fact-checker can use to conduct additional research. It displayed the event’s date and location.

With this vital information, we were able to locate the event organizer’s Facebook page. A further Google search led us to the full version of the video, which had been published by a local television station. We found that Thapa’s statement had been taken out of context in order to spread misinformation.

Thapa had argued that while populist programs receive widespread support, they are difficult to fund. To disseminate misinformation, the video, which was more than an hour long, was edited down to a 22-second clip. We could verify this by following the claim back to its source, which is an important aspect of combating misinformation.

Human intelligence and engaging the audience

While the digital age creates opportunities for the spread of fake news, misinformation and disinformation, it also creates opportunities for fact checkers to follow digital trails in order to try to find the source and the veracity of the information being shared.

The fundamental journalistic skills of fact-checking are not new, they just need applying to the digital age. Our duty, as always, is to publish accurate, fair, objective, impartial and well-sourced information, backed by well-researched context and analysis. That has not changed. And digital tools are a massive help in the pursuit of robust and critical journalism. We just need to make sure we are using those tools as well as those who propagate falsehoods do.

All newsrooms must have a fact-checking team made up of dedicated journalists who are skilled in thorough research and diligent scrutiny in order to ensure that all information shared is a true representation of news events.

Below is a list of some of the tools Nepal Check uses, and how we use them.

Human intelligence

  • Research who is behind the information being spread, try to find a trail of similar information to lead you to the source.
  • Does the fake news item carry a byline? Does that person exist? If they do contact them. By doing so you might find a new lead in your search for the truth.
  • Does what is being said stand up to historical scrutiny? Research and establish the sequence of events leading up to the claims.

Digital tools

  • Is the evidence real or manipulated or used out of context?
  • Are screenshots real or fake, can you find any evidence of the original material online or in print? If not, why not? Who could have created the screenshots and why?
  • Is video being used? Has a clip been taken out of context? As a journalist you have a duty to search for the original, watch it in its entirety, and offer a more complete picture including the situation and circumstances in which it was shot in order to add context and relevance.
  • When you find the original you will also discover the date and, possibly, the location of the original material. Do these match the details in the clip?
  • Check all social media links back to the original source to find out what other material has been shared.

Engaging your audience

  • Be totally transparent in your myth-busting research.
  • Let those who read your journalism know the steps you took and the evidence you uncovered.
  • By doing so you will help them understand more about fake news and empower them with the tools to do their own fact-checking.

 


The post Creating a strong fact-checking system first appeared on Media Helping Media.

]]>
The basics of fact-checking https://mediahelpingmedia.org/advanced/the-basics-of-fact-checking/ Thu, 10 Feb 2022 13:24:18 +0000 https://mediahelpingmedia.org/?p=2147 The growth of social media platforms has enabled people to express views and share content online, quickly and often. But not all of it is true.

The post The basics of fact-checking first appeared on Media Helping Media.

]]>
The author, Deepak Adhikari, delivering fact-checking training to a group of Nepali journalists in Kathmandu in April 2021. Image courtesy of the Centre for Media Research Nepal.
The author, Deepak Adhikari, delivering fact-checking training to a group of Nepali journalists in Kathmandu in August 2021. Image courtesy of the Centre for Media Research Nepal.

The growth of social media platforms has enabled people to express views and share content online, quickly and often. The ease of access to smartphones, data, and a wide variety of platforms has led to a deluge of online information. And not all of it is true.

Sometimes people share content without having adequate knowledge. This has been particularly evident during the Covid-19 pandemic.

When people are scared and have little knowledge about a topic they are more susceptible to spreading misinformation. That’s why misinformation spreads like a wildfire during major breaking news events.

People tend to share photos, memes or content that they find on social media without checking the veracity. In most cases, they do so with good intentions, perhaps to alert their family and friends.

But misinformation can be harmful because those reading the information tend to believe it and sometimes act on the basis of such information. And they, in turn, are likely to share it further.

In order for people to make informed decisions, they need to be provided with accurate and reliable facts. Fact-checking can play a critical role in this problematic news ecosystem.

Exposing the invisible – The Kit

Exposing the Invisible – The Kit has been developed by Tactical Tech in order to help people “develop the ability to question information that is false, find information when it is scarce and filter information when it becomes overwhelming.”

I was responsible for writing the part of The Kit that deals with fact-checking, including the following sections:

The Kit also includes experience of fact-checkers from Nepal and India.

For more information on fact-checking you can take a look at this list of articles and guides.

And if you need help understanding some of the terms used in any of these pieces you can refer to this glossary of definitions.


The post The basics of fact-checking first appeared on Media Helping Media.

]]>