Enforcement
What happens when an agent ignores the policy?
Technical Enforcement
- HTTP 403 — Block non-identifying agents
- Rate Limiting — 429 Too Many Requests via nginx/Cloudflare/WAF
- WAF Integration — Use siteai.json as config source
- Bot Detection — Cross-reference with existing services
Legal Enforcement
- eBay v. Bidder’s Edge (2000) — Violating terms = trespass
- EU AI Act (2026) — Machine-readable opt-out must be respected
- GDPR — Processing requires legal basis
- CFAA (U.S.) — Exceeding authorized access = federal crime
Pragmatic View
Major AI agents WILL respect it (legal liability + reputation). Small legitimate agents WILL (clear rules help). Malicious bots WON’T — but they don’t respect robots.txt either.