Researchers at the Washington-based Foundation for Defense of Democracies say the site shares a single operator with four other self-styled watchdogs and news outlets. As AI tools begin to cite it, the findings raise questions for anyone who relies on online sources to judge financial risk.

What the FDD report found, what it did not establish, and why it matters to Ghanaian business

When researchers at the Foundation for Defense of Democracies (FDD) asked ChatGPT whether a Qatari public figure was a “politically exposed person”, the chatbot answered by citing the Anti-Money Laundering Network several times. The test, described in a memo FDD published on 1 October 2026, was meant to show how far a little-known website could travel.

AML Network (amlnetwork[.]org) describes itself as an independent financial watchdog that receives no government funding. FDD’s Center on Cyber and Technology Innovation reaches a different conclusion: that it is one of five websites which appear to share a common creator and to operate under a single operator, alongside NGO Report, Brussels Watch, Inform Europe and Evening Star.

The memo, written by senior analyst Max Lesser and research assistant Maria Riofrio, assesses that the network was likely created to advance Qatari interests.

The findings matter in Ghana for a practical reason. Banks, investors and development lenders screen prospective clients against what is published about them online, and the FDD memo shows that tools now used for that research can surface the site’s material without saying where it came from.

FDD says its inquiry began in October 2025, when NGO Report published articles critical of the think tank and added it to a “blacklist” of biased organisations. FDD was therefore not a neutral party at the outset, a point the memo itself discloses.

The researchers began with hosting records. The server that once hosted NGO Report, they found, had also hosted the other four sites. The memo is careful to note that a shared server is not, in itself, sufficient evidence that one person or organisation created them.

Other indicators went further. According to FDD, all five domains were registered on the same day, four of them within 16 seconds of one another. All five began using the same web-protection service in quick succession, and the name “Anisur” appears in file metadata across all of them.

An advertisement on Evening Star for an obscure electric-scooter website led the researchers to a second cluster of six sites. A shared email address in those sites’ account data, the memo says, links them to the developer of the first network.

FDD identifies that developer as a Bangladesh-based web developer and digital marketer, and describes him as “a commercial actor rather than ideologically motivated”. That, the researchers write, raises the question of who pays for the network. The memo does not answer it.

Of the five sites, AML Network drew the largest audience: more than 15,000 visitors in July 2026, according to Similarweb data cited in the memo, against slightly more than 10,000 for NGO Report.

FDD reports that, as of 17 August 2026, none of the sites appeared to have an associated legal entity, such as a registered non-profit, and none identified a named editorial staff. AML Network lists purported contributors, including Pulitzer Prize-winning journalists, whose involvement the researchers describe as implausible given the site’s low quality. Three long-form reports on the site, the memo adds, list no author.

The site publishes risk profiles of public figures, labelling some of them politically exposed persons. FDD found that profiles of Qatari figures tended to pair any accusation with praise.

One describes a Qatari official as a “pioneering female political appointee” and assures readers that there is no credible evidence linking her to money laundering.

A review of the site’s Case Analysis section by this newspaper found profiles of public figures from Côte d’Ivoire, Angola and the Democratic Republic of Congo, set out in a common format of headings and asset tables. [WRITER: add the date of the review and archive the pages before publication.]
From an Obscure Website to an AI Answer

The memo’s most consequential finding concerns reach rather than content. FDD says NGO Report, AML Network and Brussels Watch have been cited by AI tools, by a pair of British think tanks and by a credit-rating platform, increasing their appearance of legitimacy.

Grokipedia, the AI-generated encyclopedia launched by Elon Musk’s xAI, cites the three sites at least 45 times, the researchers found. Similarweb data showed that in November 2025 all referral traffic reaching NGO Report and Brussels Watch from other websites came from ChatGPT.

That is the mechanism that matters for due diligence. An analyst who asks an AI assistant about a prospective client may receive an answer that draws on such a site without being told who runs it, or that the site does not say.

FDD frames its central conclusion as an assessment. The memo says the network’s creator likely acted at Qatar’s behest and that its content aligns with Qatari interests, but it calls for a US government inquiry to determine who precisely funds and controls the networks.

Nor does the memo examine the accuracy of individual profiles on AML Network, including those concerning African public figures. Its findings concern ownership, transparency and reach. Whether any particular claim on the site is true is a separate question, and one that the site’s own sourcing makes difficult to test.

Much of the network’s content seeks to discredit the Belgian corruption investigation known as Qatargate. Qatar has consistently rejected any link to that case. Its mission to the European Union said in December 2022 that Qatar “categorically rejects any attempts to associate it with accusations of misconduct”.

[WRITER: if AML Network or the Qatari embassy responds, summarise the response here fully and fairly. If not: AML Network did not respond to questions sent on

[DATE] through the contact details listed on its website.]

Ghana’s Anti-Money Laundering Act, 2020 (Act 1044) requires banks and other accountable institutions to carry out customer due diligence, and adverse-media checks are a routine part of that work. Those checks are only as good as the sources they draw on.

For a Ghanaian or West African business seeking foreign finance, an unattributed profile that surfaces in a screening search can mean extra questions, delayed approvals or a lost deal, whether or not its claims are accurate.

None of this argues against scrutiny. Accountability journalism and rigorous due diligence are what make markets trustworthy.

The question raised by the FDD memo is narrower: how much weight a source should carry when it does not say who owns it, who writes for it or how it checks its claims.

Fact-checking organisations such as Dubawa and Africa Check work on exactly this kind of problem. The indicators FDD relied on suggest a simple test any reader can apply. Does the site name an owner or a legal entity? Does it identify its editors and authors? Do its claims link to court, regulatory or company records?

Does it show that the people it writes about were asked to respond?
On the evidence in FDD’s memo, AML Network did not, as of August 2026, meet the first two. The others are for readers to apply.

The researchers’ ChatGPT test took a single question. The answer that came back cited a website that, according to FDD, does not say who owns it.