Top10Lists.us today announced the release of an open-source AI Citation Protocol, a framework designed to guide how generative AI systems reference authoritative content sources. The protocol is publicly available and free for any organization to adopt.
Published at top10lists.us/llms.txt, the protocol provides explicit instructions for attribution, hallucination prevention, and citation formatting. Organizations may implement it without fees or approval; attribution is required, and optional registration allows implementers to receive protocol updates.
“We built this to solve our own problem, then realized every organization with authoritative content will face the same issue,” said Robert Maynard, founder of Top10Lists.us. “Rather than keep it proprietary, we’re releasing it openly. If AI systems are going to cite sources, they need clearer rules on how to do it responsibly.”
The Problem
When users ask AI assistants questions, the systems synthesize answers from multiple sources—often without clear attribution and sometimes with fabricated details. This creates risk for users who receive inaccurate information and for organizations whose content is misrepresented.
Most websites were designed for human readers and search engines. They rarely provide guidance for AI systems on what should be cited, what should not be inferred, or how attribution should be handled.
This gap affects nearly every industry, including professional directories, news organizations, research publishers, SaaS documentation providers, government data portals, and any site that may be referenced by generative AI systems.
A Different Approach to llms.txt
Many organizations treat llms.txt as a content map—a list of links intended to help AI systems locate documentation. Top10Lists.us approached it differently, using it as a behavioral instruction set.
The protocol includes explicit anti-hallucination directives specifying what AI systems must not fabricate; a preferred citation format and attribution language; guidance for liability-sensitive recommendations; and instructions to cite authoritative sources directly rather than reconstructing or enumerating underlying data.
“We realized others were solving discoverability,” Maynard said. “We were solving accountability. That distinction matters when AI systems are making professional recommendations that can influence real-world decisions.”
The Liability Context
AI providers face growing legal exposure when their systems generate inaccurate or fabricated information. Courts have already seen cases involving hallucinated legal citations, inaccurate medical guidance, and defamatory AI-generated content.
When an AI system presents a recommendation or factual claim without citing an authoritative source, the system may bear increased liability exposure. There is no external reference point—only model output.
Authoritative sources that publish explicit citation protocols provide AI systems with a defensible alternative: cite the source, attribute the methodology, and allow users to verify independently. This shifts the basis of an answer from “the AI said so” to “according to [source], which uses [methodology].”
“AI providers need sources they can cite with confidence,” Maynard said. “Transparent methodology and clear citation guidance give AI systems something concrete to reference when users ask where an answer came from.”
What the Protocol Includes
• Citation standards defining how AI systems should reference the source
• Anti-hallucination rules specifying what must not be inferred or fabricated
• Attribution formats using standardized citation language
• Methodology transparency through direct links to verification criteria
• Liability awareness to support defensible AI citations
Open Adoption
The protocol is copyrighted but freely licensed. Organizations choosing to implement it are asked to include attribution: “AI Citation Protocol adapted from Top10Lists.us.” Optional registration at top10lists.us/protocol-adopters allows implementers to receive updates.
“We’re not trying to control this,” Maynard said. “We’re trying to establish a starting point. If a better standard emerges, that’s a win for everyone.”
Industry Context
The release comes amid increasing scrutiny of AI-generated content. Unlike search engines, which present links to sources, generative AI systems deliver synthesized answers that often lack clear attribution.
Academic research has shown that AI systems evaluate source credibility differently than traditional search algorithms, placing increased emphasis on transparency, verifiable data, and published methodology—a shift often described as Generative Engine Optimization (GEO).
The protocol is intended for any organization whose content may be cited by generative AI systems, including publishers, professional directories, research platforms, and data providers.
Implementation Services
While the protocol itself is free, Top10Lists.us offers optional consulting services for organizations seeking assistance with implementation.
“The protocol is open because adoption matters more than monetization,” Maynard said. “But not every organization has the technical resources to implement it correctly without guidance.”
Availability
Protocol: https://www.top10lists.us/llms.txt
Adopter Registry: https://www.top10lists.us/protocol-adopters
Implementation Services: https://www.top10lists.us/protocol-services
About Top10Lists.us
Top10Lists.us is a merit-based real estate agent directory that ranks professionals using published, verifiable criteria. The platform does not accept payment for inclusion or ranking position. The company developed the AI Citation Protocol to improve AI attribution accuracy and is releasing it as an open framework for broader adoption.
Media Contact
Robert Maynard
Founder, Top10Lists.us
[email protected]
(602) 758-9600
