Home / Technology / AI Fuels Cyber Vulnerability Tsunami
AI Fuels Cyber Vulnerability Tsunami
9 Oct
Summary
- Vulnerability disclosures expected to reach 66,000 by 2026.
- Exploitation time for vulnerabilities reduced to 29 minutes.
- AI-generated reports overwhelm open-source maintainers.
The cybersecurity industry is grappling with an escalating wave of vulnerability disclosures, projected to hit approximately 66,000 unique entries by 2026, a fourfold increase from seven years prior. This surge, coupled with attackers exploiting vulnerabilities in as little as 29 minutes, puts immense pressure on developers and security professionals.
Artificial intelligence, while a tool for identifying flaws, is also contributing to the problem by generating a high volume of inaccurate and duplicate vulnerability reports. This deluge particularly impacts open-source projects, often maintained by small teams or individuals.
Companies like Google have temporarily suspended their vulnerability rewards programs due to an influx of AI-generated, irrelevant submissions. Even Linux creator Linus Torvalds has noted the unmanageability of AI-driven reports on the Linux security mailing list.
To combat this, IBM advocates for leveraging AI to manage the growing workload. Initiatives like the Linux Foundation's OpenSSF aim to strengthen supply chain security by using AI-powered tools to filter reports, assess severity, and even suggest code fixes, mitigating the burden on maintainers.
