Cybersecurity flaw discoveries soar to record high with use of AI: data
TOKYO (Kyodo) -- The number of publicly disclosed cybersecurity vulnerabilities hit a record high in 2025 and continued rising rapidly in the first ha
In 2025, the number of publicly disclosed cybersecurity vulnerabilities reached a record high, with over 48,000 cases logged, a staggering 2.4-fold increase from 2021. The trend continued into the first half of 2026, with more than 35,000 vulnerabilities registered by June alone. The rapid rise in disclosed flaws underscores the heightened risks of cyberattacks and the critical need for prompt fixes.
Mitre, a U.S. nonprofit organization, compiles and shares cybersecurity vulnerabilities for IT vendors and researchers to analyze and address. Mitre's records rose from the 10,000s between 2017 and 2020 to 20,161 in 2021, and surged to 48,244 by 2025. This sharp increase in registered cases is largely attributed to the use of artificial intelligence, which has enabled faster identification of security flaws.
The rise in disclosed vulnerabilities is particularly alarming given the string of recent security breaches reported by Japanese companies. The Times Car car-sharing service and other firms have recently experienced unauthorized access and subsequent information leaks, highlighting the pressing need for businesses to strengthen their security measures.
When cybersecurity vulnerabilities are discovered, software patches must be applied promptly to mitigate risks of unauthorized access and exploitation by malicious actors. Failure to address the flaws in a timely manner could leave affected systems vulnerable. Mitre's database of vulnerabilities has grown exponentially since 2017, with the number of cases surpassing 20,000 by 2022.
The surge in registered vulnerabilities can be traced back to the launch of OpenAI's ChatGPT in November 2022, which sparked a wave of similar services from other companies. By 2024, the number of registered cases had climbed to 40,077, and continued to rise to 48,244 in 2025. The growing reliance on AI models, such as Anthropic's Claude Mythos, has further accelerated the identification of security flaws.
However, the proliferation of vulnerabilities is also attributable to a labor shortage in the cybersecurity industry, according to Yukimi Sota from the Japanese branch of U.S. cybersecurity firm Proofpoint Inc. This shortage is hindering many companies' ability to implement timely countermeasures to safeguard against cyber threats.
Written by urgent.news from The Mainichi's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.