What this evaluation is — and isn’t
An independent Grok assessment. Each evaluation includes an independent assessment of the account’s public profile and activity by Grok, the artificial-intelligence model provided by xAI. An account’s authenticity cannot be verified or proven with 100% certainty. Only the evidence that exists in the account and is available at the time of the check is used as the basis for the evaluation, and the findings resulting from the Grok evaluation are automated, evidence-based assessments, not verified facts.
authentix produces an automated, evidence-based estimate of how consistent a public X account’s observable signals are with typical authentic-account patterns. It is intended to be one input among many, not a final answer.
Affirmative disclosure approval is required before an evaluation. An authenticated user must affirmatively approve the then-current Privacy Policy, Terms of Service, and this Disclaimer through the Service. The approval is recorded with the disclosure versions in effect at the time. If a disclosure version changes, renewed approval is required before a subsequent evaluation.
This is not identity verification. We do not confirm who controls an account, check any government-issued ID, or perform Know Your Customer (KYC) screening.
This is not a background check. The score and decision are not, and must not be used as, a consumer report under the Fair Credit Reporting Act or similar laws. Do not use this evaluation, or any credential derived from it, as a factor in decisions about someone’s employment, credit, housing, or insurance eligibility.
This is not a determination of fact. The evaluation reflects patterns in publicly available data at a single point in time. Accounts change. Data can be incomplete, delayed, or misread by any automated system, including this one. A high score is not proof an account is genuine; a low score is not proof it isn’t.
AI Analysis is supplemental. Where available, authentix may generate an automated interpretation from the completed evaluation, deterministic signal evidence, account facts, activity aggregates, and the stored 30-day public post corpus. The AI layer does not change the AAE-1.0 score, decision, confidence or methodology, but automated AI readings, including an assessment by the Grok model from xAI, make up most of the final score shown for a check, and they may be incomplete or wrong.
A verification credential confirms that an evaluation occurred — not that its conclusion is correct. If someone shows you a credential, it means authentix ran this evaluation and reached this result at that time. It does not mean authentix, or anyone else, has confirmed the underlying facts.
If you believe an evaluation of your account, or an account you’re responsible for, is inaccurate or unfair, contact us at kellyhorak@authentix-account.com and we will look into it.
© 2026 Dovewood Capital, LLC. All rights reserved.
Automated observations
Automated observations, including those about pictures, are opinions based on public information and may be mistaken. A picture that appears to be AI-generated or edited is not evidence of wrongdoing, and the absence of such a flag does not show that a picture is real. authentix never uses facial recognition or identifies anyone from their face; a possible impersonation is flagged only from a claimed name or handle that matches a verified account together with a copied picture, banner or bio, and it is an observation, not a finding. Your AAE-1.0 result is unchanged by the evidence view, but picture readings count toward the final score.
Automated reading scores
Automated reading scores describe how public posts read to language and vision models. A low score is not evidence of fraud, impersonation or any wrongdoing, and a high score does not show that an account is genuine. The scores have not been validated against accounts of known origin. They do not change your AAE-1.0 result, but they are part of the final score.
The final score is mostly an AI opinion
Part of the final score comes from an assessment by the Grok model from xAI, and much of the rest comes from other automated readings, including an artificial-intelligence reading of the posts, so the final score is largely an artificial-intelligence opinion rather than a fixed calculation. It has not yet been validated against accounts of known origin. A conclusion such as Hard to verify or Warning signs found means the public information read in that check did not hold together or matched a specific pattern; it is not evidence of wrongdoing of any kind. A low final score is not evidence of wrongdoing, and a high final score does not show that an account is genuine. A picture read as likely or possibly AI-generated or altered can lower the final score, but it is not evidence of wrongdoing.
The final score can differ from the AAE-1.0 score, which is a fixed calculation from the account’s public profile and is the basis for verification credentials. The AAE-1.0 point scale has been revised, so AAE-1.0 scores from checks made under the earlier scale may not be directly comparable with later ones. Information that is missing or cannot be read earns no points and lowers the final score; a lower score for that reason is not evidence of wrongdoing.
Deleted and saved readings
A reading you delete cannot be recovered by you or by us, and a copy you saved on your device is not recoverable by us either: if you clear your browser data or lose the device, it is gone. A saved copy shows the reading as it was when you saved it and is not updated.
Your concerns, answered
When you say what prompted a check, the report answers each concern with Evidence found, No evidence found or Inconclusive, and the reasons behind it. Evidence found means public information in that check supports the concern; it is not evidence of fraud, impersonation or any wrongdoing. No evidence found means nothing in that check supports it, which does not prove the concern is untrue.
authentix reports may not be used, in whole or in part or even as only one factor, for employment, housing, credit, insurance, education or other eligibility decisions. Revisions to this page are listed in the legal change log.