According to Hacker News, unknown actors are conducting mass vulnerability scans while spoofing the user-agent strings of AI bots like ClaudeBot. This tactic allows them to bypass security measures that might block traditional scanners and gain unauthorized access to website vulnerabilities.

Data from Agent Analytics across 5,000+ websites reveals that overall bot traffic has decreased 2% in the last 90 days, while AI-related bot traffic, termed 'agentrification,' increased 11%. Among AI scraping agents, ClaudeBot accounts for 27% of all activity, followed by meta-externalagent at 19.5% and Amazonbot at 19.2%.

The top AI scraping agents are typically used by companies to train large language models, but now malicious actors are impersonating them to probe for security flaws. Website owners are advised to verify bot identities by checking reverse DNS and known IP ranges, tighten robots.txt rules, and monitor for unusual scanning patterns.

The impact is significant because security scanners can cause damage if they exploit found vulnerabilities. The spoofing makes it harder to distinguish legitimate AI crawlers from threats, potentially leading to breaches or data exfiltration if defensive measures are relaxed for trusted bot signatures.