Making trust visible: Building the AI discovery infrastructure publishers need

A researcher asks their AI assistant: "What's the latest evidence on this cancer treatment?" The AI responds confidently, citing three studies. One is peer-reviewed and current. One was retracted two years ago for data fabrication. One never existed at all—it's an AI hallucination. The researcher has no way to know which is which.

This isn't hypothetical. It's happening now, as AI assistants become researchers' primary discovery tools while lacking the infrastructure to distinguish verified scholarship from noise.

The discovery revolution no one asked for

Researchers are changing how they find information. Instead of searching databases, they're asking AI assistants: "What's the latest research on this treatment?" or "Find studies that contradict this finding." These AI agents are becoming the new front door to scholarly content.

But when an AI agent searches for research, it treats a rigorous study published in Nature the same as a blog post or fake paper from a predatory journal. It has no way to recognize that one went through peer review while the other was fabricated.

The infrastructure that signals trust (peer review, editorial oversight, version control, corrections) is invisible to AI systems unless publishers make it visible.

What's at stake

If publishers don't engage with AI discovery technologies, the consequences ripple far beyond revenue models. The foundation of scholarly communication begins to erode.

Research integrity becomes meaningless: Without integrity signals that AI systems can recognize, verification work becomes invisible.

Authors lose discoverability and attribution: If publishers block AI discovery, articles won't appear when colleagues ask AI assistants for relevant literature.

Copyright and intellectual property collapse: AI systems are already training on published content without permission or compensation. Without structured pathways for AI access, publishers lose control over how their content gets used, reproduced, and monetized.

Revenue models and impact metrics become unreliable: When AI agents bypass traditional discovery pathways, they bypass authentication. Publishers can't track usage, enforce subscription rights, or demonstrate value.

The knowledge divide widens: Research has far-reaching value for industry, consumers, governments, and more, and when publisher content isn’t a part of the dominant discovery pathways it loses impact.

AI development and discovery are moving fast, and mapping to future risks can help us all identify the foundations we need to build now to prepare for the future.

Controlling the path forward with a strong foundation

Technologies like the Model Context Protocol (MCP) offer publishers a way to participate in AI discovery while maintaining control.

For research integrity: Verification becomes visible to machines, when publishers can expose peer review status, retraction notices, version history, and corrections in ways AI systems can understand and prioritize.

For copyright protection: MCP provides structured pathways where publishers specify exactly what AI systems can access and under what terms. Authors' intellectual contributions remain protected while still being discoverable.

For revenue sustainability and usage intelligence: Authenticated AI discovery allows publishers to maintain subscription models, usage tracking, and revenue attribution while participating in how researchers increasingly work. Crucially, publishers can track AI-mediated usage, maintaining the analytics that drive strategic decisions.

The question isn't whether AI will reshape discovery; it already has. The question is whether publishers will help build the infrastructure that preserves scholarly integrity in this new landscape.

Learning today's infrastructure for tomorrow's standards

Whether MCP becomes the long-term standard or gives way to something else, the fundamental principles remain constant. AI systems will need ways to recognize research integrity signals; publishers will need mechanisms to authenticate access while enabling discovery. These needs won't change, even if the technical implementation evolves.

Without publisher engagement, AI systems will continue treating all content equally. Fabricated research will carry the same weight as verified science. A generation of researchers learns to rely on AI discovery without learning to evaluate sources. The scholarly communication system risks slowly losing coherence as the signals that maintain quality become computationally invisible.

Making trust visible

Publishers aren't just content providers. They're trust infrastructure. The verification, curation, and integrity work that scholarly publishing performs becomes more valuable as misinformation proliferates, but only if AI systems can recognize and privilege it.

The publishers who engage with structured AI discovery now are learning how to operate in an AI-driven discovery environment. They're establishing relationships with AI platforms while they can still influence how those integrations work. Those who wait aren't preserving the status quo. They're ceding decisions about discoverability, authentication, and quality signals to AI platforms that will build solutions with or without publisher input.

The infrastructure decisions publishers make today will determine whether scholarly publishing remains central to knowledge verification or becomes invisible middleware in a system that can't distinguish truth from fabrication.

Silverchair is actively engaged in addressing these challenges with publishers. Our solutions are released to market after thorough testing with end users and validation with clients, and we’re excited to share them more broadly with the market in the coming months.