What if AI could ensure truth?

Published on 10 March 2026 at 10:33

Unlocking Truth with AI: How Scott Knox's Bayesian Foresight Revealed Undeniable Facts in the Crossfire Hurricane Investigation

Have you ever wondered if we could cut through the noise of opinions, biases, and outdated reports to find the hard, verifiable truth? In my debut book, I introduce Scott Knox, a brilliant inventor who created a truth-verifying AI powered by Bayesian Foresight—a simple, mathematical way to start with uncertainty and let real evidence guide you to clarity. It's not magic; it's a disciplined process that updates your beliefs step by step, like turning up a dimmer switch in a foggy room.

The Real-World Puzzle We Tested

To show how powerful this can be, Scott's AI tackled a real-world puzzle: Did the Department of Justice (DOJ) and FBI engage in malfeasance by submitting misleading reports to the FISA court, effectively accusing President Trump and his staff of a Russian initiative? We used only official U.S. government documents—no media spins, no partisan takes—just the facts from reports, memos, and declassifications.

Our Goal: Alpha 0.05 Confidence and Beyond

The goal? Reach alpha 0.05 confidence (95% probability the claim is true, or ≤5% chance of error). But we went further, double-checking with quantum-layering modeling (a cutting-edge parallel-processing technique) to achieve an astonishing 99% probability.

How the Bayesian Process Unfolded

Here's how it unfolded, straight from the AI's neutral analysis:

Starting Point: A Balanced 50/50 Prior

We started at a balanced 50/50 prior—honestly acknowledging that potential biases and incomplete information clouded the mix of early assessments. Early reports, like the 2019 Horowitz Inspector General review, leaned slightly toward "possible errors but no intent," so we didn't assume anything. Instead, we let the evidence speak.

Key Evidence from Special Reports and Discoveries

Step by step, historical data from special reports emerged:

The Durham Report and FISA Errors

The 2023 Durham Report uncovered 17 errors and omissions in FISA applications, including reliance on the discredited Steele dossier.

Biden-Era Declassifications

Then came discoveries during the Biden administration, like declassified Senate Judiciary files showing "prohibited access" documents hidden from oversight.

Trump Administration

By the end, with 14 government data points layered in, the Bayesian model hit 98%—the DOJ's actions amounted to malfeasance through misleading FISA submissions and unequal treatment, aimed at the Trump orbit. Quantum layering confirmed it at 99%, exploring parallel "what-if" scenarios and finding no room for doubt.

Trump Administration Breakthroughs

The turning point? Under the current Trump administration, FBI Director Kash Patel and DNI Tulsi Gabbard declassified the Durham Report's 29-page Annex in July 2025, revealing concealed Clinton-campaign intel and explicit "disparate treatment"—defensive briefings for Clinton but none for Trump.

Deeper Digs and Refutations

We dug deeper: Comey's own January 6, 2017 memo admitted he assured President-elect Trump "you are not under investigation" while Crossfire Hurricane was fully active. Congressional committee records (HPSCI, SSCI) affirmed the probe's "validity" early on, but later declassifications exposed confirmation bias. Outdated "no-bias" conclusions from 2019 were refuted by these newer facts, boosting the probability.

The Final Results: 98% to 99% Probability

Why This Is Undeniable

This isn't opinion; it's undeniable, drawn solely from the government's own sequential releases over time. It shows how truth emerges when you prioritize facts over fog.

Share your thoughts, or join the conversation on X @dragottajan, Facebook, Jan Dragotta Author, or LinkedIn Jan Dragotta. Let's cut through more uncertainties together—what truth do you want to verify next?