2024 "Election Year" Growing "Concern About Fakes" in the World December 12, 16:19

"The election commission set up computers so that the president would win" is disinformation
that was spread during the Indonesian elections four years ago.

Next year, in 4, not only the US presidential election, but also large-scale elections will be held in Russia, India, Taiwan, Indonesia, and other parts of the world, and it is said that more than 2024 billion voters will vote.

Now, generative AI can even make it look like the candidate is talking about something that he or she is not talking about.

We interviewed the front lines of how to confront fake information.

(Broadcast on "Saturday Watch 20" on December 12)

Generative AI turns into fake information

In September, the image went viral on social media.

The image is fake and appears to have been created using generative AI. It spread in conjunction with the treated water from the Fukushima Daiichi Nuclear Power Plant.


"Since the treated water was released into the ocean, images of giant sea creatures have been circulating on social media, with some commenting that it was caused by 'contaminated water.' Generative AI is making it easier to create these fake images."

This is a fake video of Chinese President Xi Jinping speaking Japanese. It has been viewed nearly 100 million times on X.

A video created by having generative AI learn your voice, transcending language barriers.

The technology to match the spoken words and mouth movements has also evolved, and the feeling of discomfort is disappearing.

Even in Japan, fake videos of Prime Minister Kishida making obscene remarks made with generative AI have been spreading, and it is becoming more and more of a problem.

Fake information is as good as in an emergency.

Fake information, such as misinformation and disinformation, is likely to spread when high-profile events occur, such as elections, disasters, and major incidents.

The 2016 U.S. presidential election in which Mr. Trump won the election was the catalyst for the widespread attention of fake information.

In the fierce battle between Mr. Trump and Mr. Clinton, disinformation such as 'the Pope has expressed support for Mr. Trump' and 'Mr. Clinton is involved in a child prostitution organization' was widely spread on SNS.

At that time, there were many cases where articles on fake news sites were spread, but in recent years, fake information using fake images and fake videos has increased.

With generative AI allowing anyone to create fake images and videos, fakes are also "evolving."

More than 20 billion voters, a growing sense of crisis

2024 is just around the corner. Taiwan's presidential election in January, Indonesia's presidential election in February, Russia's presidential election in March, India's general election in May, and the U.S. presidential election in November. There is a possibility that there will be a general election for the House of Representatives in Japan as well.

It is estimated that more than 1 billion eligible voters worldwide will be eligible.

In December, an international conference was held in Singapore to discuss countermeasures against fakes.

Fact-checking groups and members of the media from the Asia-Pacific region spoke of a sense of urgency.

Editor-in-Chief of Audio Media, India (general election to be held by May next year)
"India is facing a flood of misinformation and disinformation, and it's an issue that everyone can get involved in."

A reporter from a Taiwanese fact-checking organization (presidential election in January next year) "
We are concerned that generative AI will have a major impact on the upcoming presidential election, and while it has become easier to create fakes, it is becoming more difficult to detect them."

An official of a Mongolian fact-checking group (prospecting a general election in June next year) "
We have seen a lot of fake information that has a negative impact on society, and next year's election will be a big challenge for us."

Indonesia: Fakes undermining elections

The conference, which was attended by 32 people from 411 organizations in 678 countries and regions, attracted attention in Indonesia, where countermeasures against fake information are a major issue.

In the 2017 Jakarta gubernatorial election, the incumbent governor of Chinese descent and Christian was defeated. During the election period, there was a large amount of disinformation spreading to slander this candidate.

In addition, in the 2019 presidential election, disinformation such as 'the election commission set up a computer so that President Joko won' spread, and it is said that distrust of the election has increased.

Under these circumstances, the Indonesian NPO "Mafindo" has a team of 16 people who conduct fact-checking every day and have published more than 1,2000 verification articles so far.

According to the group's analysis, more than 1,9 cases of fake information were identified from January to September.

More than half of them are related to politics, and we are concerned that this number will continue to increase.

In October, a fake video of President Joko giving a speech in Chinese went viral on social media.

President Joko does not speak Chinese, so it was clear that it was a fake video, but his voice was very similar and there was no discomfort in the movement of his mouth.

While there is a possibility that the sense of distance from China will be questioned, it is considered to be a fake video aimed at giving the impression that it is close to China.

Septiaj Eco Nugololo, CEO of Maffind,
said: "If someone creates and disseminates a fake video made by generative AI just before the election, we are concerned that we will not have enough time to verify and counteract it.

Fakes are more widespread than facts

Another issue was the difference in diffusion power.

When we analyzed the disinformation that 'the Election Commission's server was hacked from overseas' that spread in the 2019 presidential election, it was shared about 1,4000 times on SNS.

However, only about 700 verification articles pointing out this as a lie were shared.

The reason for this is that ▽ fact check articles appear after fake information is released, ▽ fake information stimulates emotions and tends to attract attention, while fact check articles are often text information and are difficult to attract attention.

Mr. Septiazi Eko Nugoro, CEO
of Muffindo, said, "Fact-checked articles are only read by about 5~10% of fake information, which is a big difference, and if the number of fake videos increases further in the future, the difference will become even larger."

Measures to be taken in a hurry

In response to the current situation where fact-checking alone cannot keep up, Muffindo is working on "Prebunking" that allows people to point out that disinformation and misinformation are wrong when they are exposed to it.

The idea is that if you come into contact with correct information in advance, you will not be misled by judging it as incorrect information, so to speak, like a "vaccine".

Based on the results of our analysis of the trend of fake information that has spread in past elections and recent changes in the electoral system, we are anticipating fake information that may appear in the future and taking measures in advance.

For example, in previous elections, there was a lot of disinformation about the management of the election, such as "the election results were manipulated," so we also made a video explaining the mechanism of counting the number of votes.

In addition, we are focusing on activities to improve literacy in discerning information, setting up booths at local festivals so that people can easily participate, and holding workshops in various places.

At this time, "fun" is important.

In October, we created a new card game to learn how to spot fakes.

What kind of game is it?

The player becomes a restaurateur. In this game, the owner of another restaurant who is the opponent disseminates fake information such as "poor hygiene" and "using expired ingredients", and cards such as "fact check" and "pre-bank" are used to counter it.

At that time, we will also think about what kind of specific information should be provided to counteract fake information.

If you don't have a card in hand, the restaurant's rating will decline and you will lose points, so you can learn how fake information spreads and how to distinguish it.

Fact-checking alone is not enough

Measures are also underway in Taiwan, where the presidential election is scheduled for the beginning of the year.

The Taiwan Fact-Checking Center, a private fact-checking organization, has also introduced a new tool to identify images using generative AI in preparation for fakes using generative AI.

In addition, "fact-checking alone is not enough," and in Taiwan, there are more opportunities for citizens and others to learn about the literacy of information flowing on social media.

In November, we held workshops for faculty and staff to learn how to distinguish between disinformation and misinformation.

Taiwan Fact-Checking Center He * Kei An * Kay is "Kusakanmuri ni Megumi" "
It is important for citizens to at least not share suspicious information when they come into contact with it. Fact-checking alone is not enough, and we want to spread the importance of increasing literacy."

What is required in Japan?

As there is a growing sense of caution about fakes becoming more sophisticated ahead of the global election year, what kind of measures will be required in Japan in the future?

We spoke with experts who also participated in an international conference in Singapore.

Professor Nobuyuki Okumura, Musashi
University: "Japan is a stable society with a high level of education, and there may be some delays in recognizing the need for fact-checking.
Depending on the result, the direction of the country will change after that, so we have to be most careful about the magnitude of the impact."

Daisuke Furuta, Editor-in-Chief of the Japan Fact-Checking Center: "The situation in the world (regarding fake information) and the situation in
Japan are basically getting worse. That's why we need to raise the level of fact-checking and media literacy in Japan as well."