To build a better Internet in the age of AI, we need responsible AI bot principles. Here’s our proposal.

Cloudflare’s blog proposes responsible AI bot principles to guide behaviour on the internet as AI crawlers and chatbots become more pervasive. The piece, dated 24 September 2025, outlines five starting points: public disclosure, self-identification, declared single purpose, respect for operator preferences, and acting with good intent.

It points to OpenAI’s GPTBot as an example of self-identification, noting that the bot’s user agent and IP addresses align with publicly disclosed information, enabling operators to decide whether to allow traffic. The proposals also discuss cryptographic verification via Web Bot Auth as a future standard and emphasise transparency and accountability across bot operators.

The article argues that bots should have clear purposes—such as search, AI-input, or training—to help operators decide access rights and avoid content being used for unwanted purposes. It concludes by calling for multi-stakeholder collaboration, mentioning July 1st Content Independence Day as context for Cloudflare’s balanced stance and urging ongoing discussion through forums like the IETF. According to OpenAI, these steps aim to strengthen trust and compatibility between content creators, AI companies, and operators as the internet evolves.