Technology has changed how information spreads. We believe it should also change how evidence is found, checked, compared, and made visible.
AI helps us find and organize evidence. Human editors verify it, interpret it, and decide what we publish.
At Evidence First, we use AI as part of our research and verification process. It helps us work through large amounts of information more quickly, organize evidence more clearly, and identify what may be missing.
But AI does not decide what is true.
It does not replace human judgment, source evaluation, editorial review, or final conclusions.
What AI Helps Us Do
We use AI to help:
Scan large amounts of public information
Identify factual claims
Surface relevant sources
Compare competing accounts
Summarize long documents
Translate foreign-language material
Search public records, reports, court filings, and regulatory documents
Flag contradictions, uncertainty, and missing evidence
AI can make the evidence process faster and more accessible. It can help us find the right questions to ask. But it cannot answer those questions responsibly on its own.
What Humans Decide
Human editors remain responsible for:
Verifying claims
Judging source quality
Providing context
Reviewing evidence
Weighing uncertainty
Deciding what conclusions are justified
Making the final editorial call
AI may assist the process. It does not serve as the authority.
Our Standard
For major claims, we ask:
What is the source?
Is the source primary, independent, or secondhand?
What evidence supports the claim?
What evidence complicates it?
What is missing?
What is uncertain?
Are there competing accounts?
Who benefits if people believe this?
When the evidence is incomplete, we say so.
Sometimes the most honest answer is: we do not know yet.
Why This Matters
Truth has a labor problem. Finding reliable evidence is often slow, tedious, and hard for ordinary readers to inspect.
Used well, AI can lower the cost of verification. It can help extract factual claims from speeches, articles, reports, and public records. It can help compare those claims against available evidence. It can help summarize court filings, regulatory documents, research papers, and foreign-language sources. It can also help identify contradictions, uncertainty, and gaps in the evidence.
That does not make AI the authority. It makes it a tool.
We believe AI is most valuable when it is built into a system of accountability. Evidence should be source-grounded whenever possible. Dates should be clear. Primary documents should be prioritized. Uncertainty should be stated plainly. Missing evidence should be acknowledged.
That is how we use AI at Evidence First: not as a shortcut around judgment, but as a tool to make the evidence easier to find, easier to test, and easier for readers to see.