麻豆传媒

KUMAMOTO – Generative artificial intelligence (AI) has been increasingly utilized to assist relief efforts in the wake of a powerful earthquake in Kumamoto Prefecture late last month.

At the same time, AI is believed to have been used to create fake videos and other misinformation that were spread on social media, highlighting both the advantages and drawbacks of the technology.

Misato Kaetsu, 38, who experienced the quake in the Kumamoto town of Mifune, created an online information board?using generative AI on her smartphone while at the evacuation center where she and her family had fled.?

Those affected by the quake, which registered up to 7, the highest level on 麻豆传媒 seismic intensity scale, can use the site to share local information, such as the locations of reopened stores and emergency feeding stations. The service was launched just hours after the idea was conceived.

“I decided to develop (the service) because information has not reached those who need it,” Kaetsu said. “AI could not be used when powerful earthquakes struck Kumamoto 10 years ago, but ideas can now be easily turned into reality,” she said.

More than 160,000 people have already used the service, and it has assisted those affected by the disaster in rebuilding their lives.

Staff at quake-hit local governments are also using AI.

“We have used it to create meeting minutes and explanatory materials since before the earthquake, so it has become indispensable,” an official at the digital strategy promotion division of the Kumamoto Prefectural Government said. “By reducing the workload through the use of AI, we can focus on issues that require careful handling during disaster response.”

However, AI also has its downsides in times of natural disasters.

Following the latest earthquake, a fake video of an explosion at the Aeon Mall Kumamoto shopping facility in the Kumamoto town of Kashima was spread on social media. This and other fake videos are believed to have been created with AI.

The explosion at the shopping facility, which occurred shortly after the earthquake, left seven people dead.

“In the past year or two, the accuracy of generative AI has evolved dramatically, and reliability and convenience have increased,” Isao Echizen, professor of information security at the National Institute of Informatics who is well-versed in AI technology, said. “On the other hand, there is a risk of (AI) giving wrong answers, albeit at a low probability, and being abused.”

“It’s hard to tell if an image has been generated” with AI, he added. “It will be more important than ever to check whether those who post (images) are reliable sources.”