What Colorado's complaint alleges
On November 4, 2026, Colorado Securities Commissioner Tung Chan filed a complaint against Lakewood, Colorado radiologist Michael Cole, MD, and his company, Sage Income LLC, as first reported by Radiology Business. The complaint alleges Cole and Sage issued promissory notes to 14 investors from November 2019 through January 2024, promising annual returns of 9% to 15% from a fund that was supposedly trading on an AI-driven strategy.
According to coverage of the complaint in Becker's ASC and Law360, Cole told investors Sage's algorithm could detect and "front-run" large institutional trades and had generated two to six times the returns of major stock indexes. In reality, the complaint alleges, Sage's internal records showed it was "consistently losing money" and had negative returns for the entire period from November 2019 through December 2024 — while quarterly reports sent to investors claimed the fund was beating the S&P 500, in one case by about 29.4%.
The structure, regulators allege, was Ponzi-like: new investor money and an unrelated loan were allegedly used to keep paying interest to earlier investors, and about $190,000 was allegedly transferred into Cole's personal bank account. Every one of the 14 investors reportedly lost money, with individual losses ranging from $35,000 to more than $2 million. The investor list allegedly included Cole's own parents and brother, as well as a radiologist who had supervised Cole during his residency and put in his retirement savings. Sage Income filed for bankruptcy in February 2025, and the complaint alleges Cole was never a licensed investment adviser and Sage never registered as a broker-dealer — a fact regulators say was never disclosed to the people writing checks.
Why this is not a clinical-AI story — and why it still matters to imaging
It's worth saying plainly: this complaint has nothing to do with imaging AI, diagnostic accuracy, or patient care. It's an alleged unregistered-securities fraud case built around an "AI stock-trading" pitch. The only radiology connection is that the alleged operator happens to be a radiologist, and several of the alleged victims are physicians — including radiologists.
But the mechanism is worth studying regardless of the industry it happened to occur in. An unverifiable claim — "our algorithm beats the market" — was given a credibility boost by attaching the word "AI" to it, then reinforced with fabricated performance reports. That is precisely the pattern buyers of clinical AI need to be able to spot, because the same shortcut — "it's AI, so it must be sophisticated and accurate" — is exactly what makes AI-washing in healthcare technology work.
AI-washing: the pattern underneath the headline
"AI-washing" is using the word "AI" to make a claim sound more rigorous or more validated than it actually is — without the regulatory clearance, independent evidence, or disclosed limitations that would let an outside party check the claim. It shows up in investment pitches, consumer products, and, occasionally, in healthcare technology marketing. The tell isn't the presence of AI — plenty of AI tools genuinely work. The tell is an unverifiable performance claim paired with the absence of anything a buyer could independently check.
In the Sage Income complaint, nothing about the "algorithm" was checkable: no regulator oversaw the trading claims, no independent auditor verified the reported returns, and the business wasn't even registered as the type of entity that is supposed to make these pitches. In clinical AI, the equivalent unchecked claims look like an accuracy percentage with no clearance number behind it, a case study instead of a peer-reviewed study, or a "fully autonomous" pitch with no visible human review.
Three checkable signals for any clinical AI vendor
A buyer evaluating a radiology AI reporting vendor doesn't need to take an "AI-powered" claim on faith. Three things can be verified independently of the vendor's own marketing:
| Signal | Unverified AI claim (red flag) | Evidence-backed clinical AI (green flag) |
|---|---|---|
| Regulatory status | "AI-powered" with no clearance number, or a claim that doesn't match what the clearance actually covers | A specific FDA clearance (510(k) or De Novo) for the exact claimed use, checkable in FDA's public AI/ML-enabled device database |
| Performance evidence | Vendor-reported numbers, internal white papers, or testimonials with no outside review | Peer-reviewed validation published in an independent journal, tested on data outside the vendor's own training set |
| Human oversight | "Fully autonomous" pitch, or no clear answer to "who is accountable if this is wrong?" | A named, visible radiologist-in-the-loop step before any report reaches a patient's chart |
None of these three checks is exotic. They're the same category of due diligence a financial regulator applies to an investment pitch: is the activity registered and licensed, is the performance claim independently auditable, and is there a disclosed, accountable human in the chain rather than a black box making unsupervised decisions. The Sage Income complaint alleges all three were missing. A radiology group evaluating an AI reporting vendor can ask for all three up front.
Where xAID fits
This case is a reminder that "AI" is a label, not a credential — the credential is the clearance, the published validation, and the visible human accountability behind it. AI CT reporting built on foundation models is designed around exactly those three checks: a documented regulatory pathway for the claimed use, published validation evidence, and an in-house radiologist who reviews every preliminary report before it goes out ready-to-sign to the client's reading radiologist. The goal isn't to ask buyers to trust the word "AI" — it's to give them something to check.
Frequently asked questions
What is the radiologist AI investment scam case about?
Colorado's securities commissioner, Tung Chan, filed a complaint on November 4, 2026 against Lakewood radiologist Michael Cole, MD, and his company Sage Income LLC. The complaint alleges Cole raised about $5.2 million from 14 investors — including his parents, his brother, and a radiologist who had supervised his residency — by issuing promissory notes promising 9% to 15% returns from an 'AI' stock-trading algorithm. The complaint describes it as a Ponzi-like scheme: Sage allegedly had negative returns from November 2019 through December 2024 while sending investors quarterly reports claiming it was beating the S&P 500, in one case by about 29.4%. Sage filed for bankruptcy in February 2025.
Is this a clinical AI story or a financial fraud story?
It's financial fraud, not a clinical-AI failure. The complaint concerns an unregistered securities offering and an AI stock-trading claim, not an AI tool used for imaging diagnosis. The only radiology connection is that the alleged operator is a radiologist and several of his alleged victims were physicians, including radiologists. It is still relevant to imaging because it shows how easily the word 'AI' can be attached to an unverified claim to make it sound credible — a dynamic buyers of clinical AI should also watch for.
What is AI-washing in healthcare?
AI-washing is labeling a product or pitch 'AI-powered' to borrow credibility the underlying technology hasn't earned — without the regulatory clearance, independent validation, or disclosed limitations that would let a buyer check the claim. In the Colorado case, the alleged 'AI trading algorithm' was never shown to work and the business wasn't registered as a broker-dealer or investment adviser. In clinical AI, the equivalent red flags are accuracy claims with no FDA clearance number, no peer-reviewed validation, and no disclosed human-review step.
What should buyers check before trusting a radiology AI vendor's claims?
Three checkable signals separate evidence-backed clinical AI from unverified AI claims: (1) a real FDA clearance or authorization — searchable in FDA's public AI/ML-enabled device list, tied to the specific claimed use, not just a general "AI-powered" label; (2) a radiologist-in-the-loop workflow, where a human reviews and is accountable for every report rather than the AI operating autonomously; (3) published, peer-reviewed validation on an external dataset, not only a vendor's own white paper or marketing claim.
Does FDA clearance alone guarantee an AI radiology tool works as claimed?
No — clearance confirms a tool met a defined regulatory bar for a specific claimed use, not that it performs exactly as marketed in every setting. That's why clearance is one checkable signal among several, alongside independent peer-reviewed validation and a visible human-review step, rather than a single box that substitutes for due diligence.
Source: Radiology Business, reporting on the complaint filed by Colorado Securities Commissioner Tung Chan against Michael Cole, MD, and Sage Income LLC (Nov. 4, 2026); additional detail from Becker's ASC and Law360. Figures are rounded as reported and reflect allegations in a civil complaint, not a court finding.