Moldova records platform referrals. The public cannot yet audit them

The 2025 campaign left outsiders guessing

The transparency problem predates the Commission’s latest report. In its final election assessment, ODIHR said Moldovan authorities sent around 1,000 takedown requests to platforms during the September 2025 parliamentary campaign. Most were not acted on. ODIHR could not establish why authorities had flagged the content or why platforms had rejected the requests.

That gap prevented outside scrutiny in both directions. Without the referral grounds, voters could not assess whether authorities were applying clear criteria. Without platform responses, they could not distinguish a rejected weak case from delayed action, inconsistent moderation or a platform’s refusal to enforce its own rules.

Responsibility was also divided. ODIHR recorded that the Central Electoral Commission, the Security and Intelligence Service and police dealt with TikTok. STRATCOM handled Meta and Google, while the Audiovisual Council covered YouTube news videos. A public account should therefore identify the referring institution. A single national total would hide differences among agencies and platforms.

STRATCOM already keeps most of the necessary fields

STRATCOM’s written referral methodology draws an important boundary. The Centre submits reasoned referrals, while platforms alone decide whether to remove, restrict, label or leave content unchanged. These are not state removal orders.

The methodology also excludes ordinary criticism, satire, journalism and legitimate political opinion unless the material involves fabrication, impersonation, deceptive practices or coordinated inauthentic behaviour. For each referral, STRATCOM records the date, platform, object, grounds, requesting authority where applicable, and the platform’s response or non-response. It may publish aggregated figures, but the document does not require a regular report or specify that the date of the platform’s response must be recorded.

“Moldova already records the grounds and platform response for each referral internally. The missing step is an aggregate public record.”

Written safeguards matter, although they cannot show how consistently officials apply them. Regular disclosure would allow journalists, researchers and parliamentarians to compare the methodology with actual referral patterns without exposing personal data or operational details.

One Gagauzia case shows what remains unknown

A 30 October STRATCOM case illustrates the problem. Three small Facebook pages promoted an unauthorised electoral survey in Gagauzia through about 24 Romanian- and Russian-language ads. The organiser and funder were undeclared. STRATCOM could not establish how many people completed the survey or whether it affected political preferences.

The public account describes the suspected manipulation, but it does not say whether STRATCOM referred the ads to Meta, when any referral occurred or what Meta did. Russian-language elements and a Volga identifier do not prove Russian state direction. ODIHR documented Russian interference during Moldova’s 2025 parliamentary campaign, yet that wider record cannot substitute for evidence about this particular operation.

The fifth EU-Moldova political and security dialogue on 5 November addressed cyberattacks and disinformation campaigns, according to the Moldovan foreign ministry. Its public release announced no platform-referral reporting mechanism. The Commission separately noted that Moldova still needs enforcement capacity and a legal basis for applying the EU’s platform rules as of accession.

A ledger can disclose results without amplifying content

A useful public record should identify the referring institution, platform, referral category and stated basis. It should then report whether the platform removed, restricted, rejected or ignored the material. To measure response times, STRATCOM would also need to record and publish the date of each platform response alongside the referral date. Corrections and reversed referrals should appear as separate categories because they test whether internal review works.

That ledger is the direct answer to Moldova’s accountability gap. It would let the public test proportionality by comparing grounds with outcomes and effectiveness by tracking action across platforms. Added response timestamps would make timeliness measurable. Publication of those fields on a regular schedule, or a binding rule producing equivalent data, would change the present judgment. Until then, Moldova can document much of its intervention process internally, but citizens cannot independently assess how the referral system performs.