The crisis of journalism: How to tackle fake images and disinformationApril 4 at 14:19

In late March, when news emerged that former US President Trump might be arrested, tensions were rising, and there is an image that has spread around the world.

It is a fake image of "former President Trump being seized by police".

The fact that even such images can be easily created using AI = artificial intelligence has made journalists working on "fact-checking" in the world media express a sense of crisis.

(Junya Yabuuchi, Department of Science and Culture / Chihiro Tanaka, World News Department)

The threat is "real, fast and anyone can do it"

"What happens to the news if you don't believe what you see?"

was held in London at the end of March at a meeting called "TNI" = "Initiative for Reliable News". About 3 journalists from 14 countries gathered at the annual meeting of a liaison organization led by the BBC, the UK's public broadcaster, and includes international media and IT platforms.

We were the first Japan media outlet to participate.

In the session, we will introduce each other's countermeasures against disinformation and learn from each other.

Broadcasters from all over Europe worked together to carefully collect publicly available data and follow the movements of children taken from Ukraine by Russia.

How the former president's camp responded to the attack on the media in Brazil's presidential election.

In many of the sessions, participants touched on the threat of AI.

The fake images introduced at the beginning of this article in which former President Trump is seized are said to have been created as a joke by the founder of the international investigative reporting group Bellingcat, but such images can be easily created by creating a "generative AI" that generates images when commands are entered.

AI experts


who have advised the BBC and others in a session titled "Generative AI What happens to news when we can't believe what we see?"
He pointed out that it is completely different from the past.

Recently, in addition to fake images of former President Trump, fake images of the Pope in a down coat of a Spanish luxury brand went viral on social media.

Even in Japan, when flooding occurred in Shizuoka Prefecture due to heavy rains caused by Typhoon No. 9 last September, a fake image made by AI was posted on Twitter as an image of the flood.

In fact, if you look closely, there is something unnatural about the image, but if you are in a hurry, you may retweet it as real.

When anxiety is spreading due to disasters or incidents, they become uncalm and easily spread images and information of dubious origin. Disinformation that is spread with some intent in the form of taking advantage of such emotions is becoming more sophisticated every year.

Technical measures are "difficult"

At the meeting in London, while a strong sense of crisis about AI was expressed, it seemed that technical measures were lacking in decisiveness.

There was also discussion of various technologies, such as using blockchain technology that can track the source of data to authenticate images and using AI to detect disinformation and fake images.

However, new measures are always overwritten by new technologies, and "technology can only be used for about six weeks." Even at the venue, there were voices saying, "It's just a cat-and-mouse situation."

In response to the ever-increasing threat of disinformation, including the evolution of AI, it is necessary to simultaneously promote measures such as technological development, the development of human resources to detect disinformation and respond appropriately, and literacy education on the receiving side.

Countermeasures against disinformation that spreads unverified on social media have become an issue in media around the world.

Specialized departments and "SNS reporters"

The BBC has created a new department to combat disinformation.

The department, named the Forensic Journalism Hub, began operations on April 4.

"Forensic" means "forensics," and it is said to be a group of 3 people, including a group that verifies disinformation, so-called fact checking, an investigative reporting group called "OSINT" that finds facts from open data, a person in charge of SNS, and a video department.

Placed within sight of the BBC's news centre, it allows more people to participate openly and reminds them of their commitment to combating disinformation.

Mariana Spring, the BBC's first "disinformation and social media reporter," is 27 years old.

He points out disinformation, interviews people who are spreading disinformation, and disseminates it on BBC programs, social media, and podcasts.

What Ms. Spring has been tackling is the problem of "disaster trolling," the act of ridiculing disaster victims on the Internet and attracting attention.

Regarding the 2017 terrorist bombing at a live venue in Manchester, he directly hit a man who was spreading false information that "no terrorist incident has occurred, the dead are alive, and those who are injured are acting," and conveyed the thoughts of the victims.

Some people have attacked Ms. Spring on social media, but she says that she not only overwrites disinformation with correct information, but also includes the background of why people who spread disinformation spread it and why people who believe in disinformation believe it.

Mariana Spring
: "We want to understand why people get caught up in disinformation and why they believe it. I think a lot of information isn't being trusted, and other issues people are experiencing, may also play a role, and we need to know that."

"Show the process and trust it"

As social media overflows with information and trust in traditional media declines, there is a need for changes in the way the media is communicated.

The word "transparency" was repeatedly heard at the meeting.

Ross Atkins, a BBC anchor, said during the meeting that in order for the media to gain trust, it is necessary to provide information that viewers and readers can verify, and it should be in a way that says, "We know this much now, and we don't know any more."

An example of this is a case where the news deliberately reported that "the facts have not been confirmed yet."

Last year, when China ended its zero-corona policy and the number of infections was increasing explosively, the authorities did not provide sufficient data.

However, on Chinese social media, videos of patients rushing to hospitals are posted one after another. The BBC couldn't confirm whether the footage was correct.

However, it is true that they are posted one after another.

After discussing with the staff, they decided to tell what was happening in the news after saying that "the facts have not been confirmed yet."

We dare to show information that the news organization has not been able to confirm the facts, which is fundamental for news organizations. It raised the question of whether trust should be gained by transparently showing how information was obtained and how it was communicated.

Ross Atkins:
"There was a big debate among the staff about whether we should provide information that has not been fact-checked, and we thought that we could increase trust by showing the process as well as conveying useful information to viewers. As misinformation spreads so quickly, we need to change the way we present information as a media."

What is needed to combat disinformation?

Until a few years ago, the BBC and Reuters had expressed the opinion that journalists were in the job of checking the facts and writing articles, and that it was the job of journalism to combat disinformation.

However, in particular, the "Brexit" in which the United Kingdom decided to leave the EU in 2016, and the election of President Trump in the United States, the opinion that the decision may have been made after spreading information that is not necessarily true has strengthened.

Reflecting on the fact that the media may not have adequately conveyed the facts in response to the developments at that time, the movement to promote measures against disinformation in the Western media is accelerating.

As the threat to fake images and texts generated by AI is increasing, the three initiatives necessary to combat disinformation were conveyed at the meeting:
▽ "collaboration" Collaborate
not alone but multiple media ▽ "Technology" Develop
technology to detect disinformation ▽ "Narrative" Rather than counteracting a large amount of disinformation one by one, these were three ways

of telling the story by fact-checking it from a bird's-eye view of the entire story, including the background of the disinformation.

NHK is also trying to further strengthen its movement by issuing programs and news that point out the spread of disinformation and misinformation in the Internet space. In the midst of rampant disinformation, how should we position measures against misinformation and in what form will we provide information? I felt that finding was now required.

Science and Culture Department Desk
Junya YabuuchiAfter working as a correspondent in New York under the
Trump administration in 1996,
I was in charge of the new coronavirus for three years as a medical reporting deskI would like to play a role in
preventing the spread of unfounded medical information and disinformation.

While exchanging information with media representatives in various countries, such as the
international broadcaster World News Chihiro

Tanaka joined the BBC in 1997, we are strongly aware of the importance of countermeasures against disinformation.