Japan said on Friday it will set up a team to simulate cyberattacks using artificial intelligence to identify and address critical vulnerabilities in PHP, a programming language used by nearly 70 percent of the world’s websites.
“An attack can succeed by getting through just one hole ... for that reason it’s important we shift our mindset and think from the perspective of the attacker,” Japanese Economy Minister Ryosei Akazawa said while announcing the initiative on Friday.
“Defenders also need to make use of AI and respond proactively, staying one step ahead.”
Major Japanese companies and government agencies temporarily suspended online services in response to the PHP vulnerabilities.
Japanese mobile carrier SoftBank said on Friday it was suspending its emergency alert system for mobile users.
That means customers will not receive early warnings for earthquakes, tsunami alerts, or disaster and evacuation information as the system undergoes maintenance, SoftBank said.
It also temporarily took down the homepages of Y!mobile and LINEMO, two of its mobile phone brands, during the emergency maintenance.
“An attack can succeed by getting through just one hole ... for that reason it’s important we shift our mindset and think from the perspective of the attacker,” Japanese industry minister Ryosei Akazawa said on Friday.
“Defenders also need to make use of AI and respond proactively, staying one step ahead.”
Financial services firm Nomura Asset Management said it began restoring its website on Thursday after precautionary security checks, adding that it found no evidence of cyberattacks or information leaks.
Japan’s Ministry of Finance also began restoring suspended online content after implementing security measures, while paint manufacturer Nippon Paint Holdings announced on Friday that it had completed emergency maintenance to address the vulnerability.
Attackers likely target Japanese businesses for their thorough records, which AI has made easier to access because they no longer face a language barrier, said Eva Chen, CEO of Trend Micro.
Japan has been too comfortable with its existing infrastructure, which it felt was safe enough, she said.
All Nippon Airways (ANA) temporarily suspended its pocket app and the ability to issue boarding passes to Apple Wallet through its website and app on Thursday as a “preventative measure to proactively prevent potential security risks.
“Please note that this matter does not indicate that our company has been subjected to a cyberattack or that any information has been leaked,” the airline said.
The U.S. National Security Agency (NSA) said on Oct. 21 that organizations using PHP must update their systems immediately, warning that the vulnerabilities could let hackers take control of affected web servers.
The agency said it is seeking out the “malicious cyber actors” that may exploit these vulnerabilities, collectively nicknamed “PHP2Shell.”
Tech Bosses Sound Alarm
Jason Clinton, Anthropic’s deputy chief information security officer, said the cyber chaos is just beginning.
“As bad as things have gotten in the past four days, they are going to get a lot worse before they get better,” Clinton said in an Oct. 23 post on X.
Government and commercial websites in the West have been temporarily shut down as operators scrambled to protect systems from possible attacks.
Thirty-two UK government services were unavailable, and four were partially unavailable amid the PHP cybersecurity issue, UK tech minister Liz Lloyd said in the House of Lords on Oct. 22.
French prefecture websites were also hit.
In the Netherlands, several government websites, including Overheid.nl, MijnOverheid.nl, DigiD, and its data protection authority, were taken offline.
In Denmark, the country’s national school communications platform, Aula, shut down on Oct. 22 because it was “addressing a potential technical security risk.”
OpenAI said that it helped to identify one of the vulnerabilities using its GPT‑5.6 Sol chatbot model.
In an Oct. 21 blog post, it said Calif, a California-based cybersecurity and AI-security company, identified the critical flaw.
Google-owned cloud cybersecurity company Wiz also contributed, saying it used AI bots such as Wiz Atlas, Gemini 4 Argon, and GPT-5.6 Sol throughout this research.
“These findings rank among the most severe vulnerabilities discovered in PHP, and arguably in software history,” Wiz said in an Oct. 21 statement.
“[The] underlying flaws had been present for decades, but demonstrating their exploitability required an exceptionally complex chain of steps, at the edge of what human researchers could practically achieve.
“Frontier AI models substantially reduced that barrier, turning a longstanding flaw into an exploit.”
Reuters contributed to this report.
English (US) · 
Turkish (TR) · 