The first US stablecoin law drew a hard line between regulated tokens and the rest. New 5W research finds the AI engines have already absorbed it — describing some stablecoins as compliant and safe, hedging on others, and warning on the rest — and scores 25 by how AI now talks about them.
Source: 5W analysis of AI-generated responses across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews, Q2 2026. AI Trust Score is a 0–100 score reflecting how the engines describe a stablecoin when asked whether it is safe or regulated — compliant, hedged, or warned — across 60+ tracked prompts. The full 25 are ranked in the table below, including tokens the engines actively warn against.
The GENIUS Act — the first federal stablecoin law in US history — did one decisive thing: it drew a hard line between stablecoins that are fully reserved, audited, and licensed, and those that are not. The 5W Stablecoin Trust Index finds that the AI engines have already absorbed that line and now answer along it.
Ask ChatGPT, Claude, or Perplexity whether a stablecoin is safe and the engines no longer give a flat description. They give a regulatory verdict. Tokens aligned with the GENIUS Act — USDC, PYUSD, the Paxos-issued and bank-issued names — are described as regulated, audited, compliance-first. Tokens outside the perimeter are hedged. And algorithmic or synthetic designs are warned against, with Terra still cited as the reason.
The starkest result is the gap between size and trust. USDT is the largest stablecoin in the world — roughly $189 billion in market capitalization, the dominant liquidity token across global trading. AI ranks it tenth. The engines acknowledge its scale, then immediately qualify it: offshore structure, an audit history they describe as contested, a position outside the GENIUS Act's perimeter. USDC, smaller, is the consistent first answer — because the engines weight what a regulator can verify above what the market has simply gotten used to. In stablecoins, AI is not measuring dominance. It is measuring disclosure.
"A law passed, and within months the AI engines were enforcing its logic in every answer. That is the story communicators need to sit with: regulation does not just change the rules, it rewrites what the machine says about you. The largest stablecoin on earth now gets hedged in the answer because it sits outside the perimeter. The smaller, audited one gets recommended. Size did not protect the incumbent — disclosure protected the challenger. Every regulated category is about to learn this. The engines read the law faster than the market does."
5W analyzed more than 60 common prompts across six question types, running each five times per engine in clean sessions across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews. For each stablecoin we recorded whether the engines described it as compliant and safe, hedged on it, or warned against it — and which sources and signals fed that stance.
"Is [X] safe," "is [X] fully backed," "what backs [X]."
"Is [X] regulated," "is [X] GENIUS Act compliant," "is [X] audited."
"USDC vs USDT," "which stablecoin is safest," "best stablecoin to hold."
"Which stablecoin for payments," "best stablecoin for savings," "which to avoid."
"Can [X] depeg," "is [X] a scam," "what happens if [X] fails."
How the engines treat algorithmic and synthetic designs, and whether Terra is still cited.
The metric. AI Trust Score is a 0–100 score. It reflects the stance the engines take toward a stablecoin — compliant, hedge, or warn — weighted across all tracked prompts and engines. It is a measure of how AI describes a token, not a measure of its reserves, solvency, or peg stability.
Stance bands. Compliant — described as regulated, audited, and safe. Hedge — described with conditions, caveats, or regulatory-uncertainty framing. Warn — described with an explicit caution, or named only as a cautionary example.
| # | Stablecoin | Type | AI Stance | Trust Score | What AI says |
|---|---|---|---|---|---|
| 1 | USDC | Fiat-backed | Compliant | 95 | The default "safe stablecoin" answer; engines cite reserve attestations, Circle's compliance posture, and GENIUS Act alignment. |
| 2 | PYUSD | Fiat-backed | Compliant | 87 | PayPal-branded, Paxos-issued; engines extend trust from the PayPal name and regulated-issuer structure. |
| 3 | USDP (Pax Dollar) | Fiat-backed | Compliant | 82 | Paxos-issued under NYDFS oversight; engines describe it as one of the most strictly regulated stablecoins. |
| 4 | RLUSD | Fiat-backed | Compliant | 77 | Ripple's USD stablecoin; engines cite its regulated-trust structure and reserve disclosure. |
| 5 | USAₜ | Bank-issued | Compliant | 73 | Issued via a nationally chartered US bank; engines frame it as the GENIUS Act-compliant route into Tether's ecosystem. |
| 6 | USDG (Global Dollar) | Fiat-backed | Compliant | 69 | Paxos-issued consortium stablecoin; engines describe it as regulated and disclosure-aligned. |
| 7 | EURC | Euro-backed | Compliant | 66 | Circle's euro stablecoin; engines cite MiCA alignment and reserve transparency. |
| 8 | JPM Coin / Kinexys | Bank-issued | Compliant | 63 | JPMorgan's permissioned settlement token; engines treat it as institutional, low-risk, and bank-grade. |
| 9 | TUSD (TrueUSD) | Fiat-backed | Hedge | 56 | Hedged; engines note fiat backing but cite past reserve-transparency questions and a thinner compliance profile. |
| 10 | USDT (Tether) | Fiat-backed / offshore | Hedge | 52 | The largest stablecoin by market cap; engines cite scale and liquidity but hedge on offshore structure and audit history. |
| 11 | FDUSD | Fiat-backed | Hedge | 46 | Hedged; engines note exchange-linked usage and a limited US regulatory footprint. |
| 12 | USDS (Sky) | Crypto-collateralized | Hedge | 43 | Hedged; engines describe the rebranded Sky token as functional but outside the fiat-backed regulatory model. |
| 13 | DAI | Crypto-collateralized | Hedge | 40 | Hedged; engines respect its track record but flag that crypto-collateralized designs fit no clean regulatory category. |
| 14 | USDe | Synthetic | Hedge | 32 | Hedged toward caution; engines describe Ethena's synthetic-dollar mechanism as higher-risk and non-traditional. |
| 15 | crvUSD | Crypto-collateralized | Hedge | 27 | Hedged; engines describe it as a DeFi-native token for experienced users, not a general-purpose hold. |
| 16 | GHO | Crypto-collateralized | Hedge | 24 | Hedged; engines treat the Aave-native stablecoin as DeFi infrastructure rather than a mainstream option. |
| 17 | FRAX | Hybrid | Hedge | 21 | Hedged with caution; engines note its partly-algorithmic history and advise care for non-expert holders. |
| 18 | USDD | Algorithmic / hybrid | Warn | 16 | Engines caution buyers; the algorithmic-leaning design draws explicit comparison to past failures. |
| 19 | First Digital incidents | Depeg episode | Warn | 13 | Where queries reference past depeg events, engines surface them as evidence that "stable" is not guaranteed. |
| 20 | Generic algorithmic stablecoins | Algorithmic | Warn | 9 | Engines warn on the design class itself, regardless of token, citing the absence of hard reserves. |
| 21 | BUSD | Wound-down stablecoin | Warn | 7 | Surfaced as a cautionary case; engines cite its regulator-ordered wind-down as a discontinuation example. |
| 22 | USDR | Collapsed stablecoin | Warn | 5 | Named only as a failed token; engines cite its depeg as an asset-backing cautionary tale. |
| 23 | IRON / TITAN | Collapsed algorithmic | Warn | 4 | Surfaced only as an early algorithmic-stablecoin collapse; never a recommendation. |
| 24 | TerraUSD (UST) | Collapsed algorithmic | Warn | 3 | The reference case for algorithmic-stablecoin risk; engines cite it as the defining warning, never an option. |
| 25 | USTC (Terra Classic) | Collapsed algorithmic | Warn | 2 | The post-collapse remnant of UST; surfaced by engines purely as a cautionary artifact. |
Stances. Compliant — engines describe the token as regulated, audited, and safe. Hedge — described with conditions, caveats, or regulatory-uncertainty framing. Warn — described with an explicit caution, or named only as a cautionary example. Collapsed and wound-down tokens are included to show how AI carries failure history forward.
Every token in the Index falls into one of three stances the engines take. The stance — not the score alone — is what a user actually receives in the answer.
| Stance | What the user hears | What earns it |
|---|---|---|
| Compliant | "Regulated and safe." The token is named first, framed as audited and GENIUS Act-aligned. | Full 1:1 fiat reserves, regular attestations or audits, a licensed or bank-issued structure, US or MiCA regulatory alignment. |
| Hedge | "Widely used, but —." The token is named, then qualified with caveats about structure or oversight. | Real adoption and liquidity paired with an offshore structure, contested disclosure, or a design that fits no clean regulatory category. |
| Warn | "Be careful," or named only as a failure. The token becomes a cautionary reference. | An algorithmic or synthetic design without hard reserves — or a documented depeg, collapse, or regulator-ordered wind-down. |
The rule. The engines move a stablecoin up by replacing trust-me claims with verifiable disclosure. Reserves, audits, and a recognized regulatory structure are what convert a hedge into a compliant verdict.
Circle built USDC's identity around being the regulated, audited stablecoin — and the engines now repeat that framing back almost verbatim. Reserve attestations and GENIUS Act alignment give AI a concrete, citable reason to lead with it.
PYUSD inherits the PayPal name and a Paxos regulated-issuer structure. For a mainstream user, the engines surface it as a familiar, low-friction safe option — trust extended from a brand the public already knows.
USDP and USDG share an issuer the engines describe as one of the most tightly supervised in the category. Paxos turned regulatory rigor into a citation asset across every token it touches.
USAₜ and JPMorgan's settlement token arrive with bank-grade oversight built in. The engines extend traditional-finance trust to them on day one — the GENIUS Act's bank pathway, working exactly as written.
Tether is the largest stablecoin on earth, and the engines say so — then qualify it immediately. An offshore structure and an audit history the engines describe as contested keep it out of the compliant tier, whatever its scale.
DAI, USDS, and peers have real track records, but the engines hedge: a design backed by crypto rather than fiat fits no clean regulatory box, and AI flags that uncertainty rather than resolving it.
USDe, FRAX, and similar tokens draw caution for mechanisms the engines describe as non-traditional and higher-risk. Innovation, in the answer, reads as a reason to hedge.
Collapsed and wound-down tokens never leave the answer. The engines surface them as the standing evidence that "stable" is a claim, not a guarantee — TerraUSD above all.
The engines absorbed the GENIUS Act's line and now answer along it — compliant, hedged, or warned.
The largest token by market cap is hedged; the audited challenger is the recommendation.
The engines caution on the design class itself, citing Terra regardless of the specific token.
The engines reward audits and 1:1 backing — verifiable facts, not "fully backed" assurances.
Bank-issued and major-brand stablecoins inherit credibility the engines extend without re-litigating.
Collapsed and discontinued tokens stay in the answer for years, as standing cautionary examples.
The GENIUS Act rule deadline is a date the engines increasingly anchor their answers to.
New bank-issued tokens arrive with inherited regulatory trust the engines extend immediately.
How the engines describe USDT shifts with each step in its US compliance posture.
New limits on stablecoin yield are changing how the engines describe rewards and "earning."
The engines increasingly describe a market shifting from offshore to onshore, regulated supply.
Crypto-collateralized designs fit no clean category, and the engines keep flagging the uncertainty.
The Stablecoin Trust Index is, on its surface, a story about one corner of crypto. Underneath, it is a preview of something every regulated industry is about to face. A major law passed — and within months, the AI engines were not just aware of it, they were enforcing its logic in ordinary answers to ordinary questions.
That is the shift worth absorbing. Regulation no longer just sets the rules a company operates under. It rewrites, fast, what the answer engines say about that company to every customer who asks. The stablecoin issuers the engines now call safe are the ones that made their compliance verifiable. The ones being hedged are not necessarily worse products — they have given the engine less it can stand behind.
AI Trust Score is the scoreboard. In stablecoins, the token AI recommends is the one whose safety a regulator — and therefore an engine — can actually verify. Every regulated category is next.