Cloudflare has publicly accused AI search company Perplexity of using stealth crawling techniques to bypass website restrictions such as robots.txt directives and network blocks. According to Cloudflare, Perplexity employs undeclared user agents, rotating IP addresses, and obscures its crawling identity to evade detection and access content from sites that explicitly block AI scraping. Cloudflare's August 4 blog post detailed that Perplexity repeatedly modifies its web-crawling bots to circumvent data-scraping defenses on millions of third-party websites. This has raised ethical and cybersecurity concerns within the technology community. Perplexity has responded by disputing Cloudflare's claims, arguing that the accusations are based on errors and misunderstandings, and clarifying that its AI assistant operates on real-time, user-driven requests rather than traditional web crawling. The dispute highlights ongoing tensions between AI companies and web infrastructure providers over data access and internet trust.