---

Discussion

Ann Tofado:

Graphika did not describe one tidy campaign with one message. It identified three activity sets: a likely successor to a previously removed Spamouflage-linked account; four accounts with Spamouflage histories amplifying a conservative-styled persona; and a large account presenting itself as the Philippine culture blog Coolture that later shifted toward pro-China and divisive U.S. political content. The common thread was not support for one candidate, but the use of apparently domestic voices to pressure democratic legitimacy.

Red Velhouse:

So the objective may be less “vote for this person” than “believe the whole system is corrupt.” Is that fair, or does it go beyond the evidence?

Ann Tofado:

It would go too far to call that a proven operational objective. But the content questioned election integrity, attacked institutions, and circulated messages both supporting and criticizing Donald Trump. That inconsistency matters. Contradictory messages can still produce the same effect: more suspicion, hostility, and doubt that a political outcome is legitimate.

Omar Seidren:

Contradiction is not necessarily a technical flaw. A persuasion campaign optimizes for a consistent audience and message; an agitation system targets different communities with whatever grievance is available. It can be pro-Trump in one post, anti-Trump in another, and broadly anti-institutional throughout. That is less a traditional campaign than a content system looking for emotional traction.

Red Velhouse:

Omar, what makes this more than ordinary anonymous political speech? People use pseudonyms online all the time.

Omar Seidren:

The distinction is behavioral and networked, not simply linguistic. Researchers examine account histories, coordinated amplification, repeated links, sudden identity changes, and whether engagement comes from the same network. Graphika also pointed to full-width punctuation and other artifacts consistent with people typing English on Chinese-language keyboards. That is an attribution clue, not proof of nationality. Combined with coordinated behavior and repurposed identities, it becomes more informative than any single odd comma.

Ann Tofado:

The identity construction is politically important. A post labeled foreign propaganda is received differently from one that appears to come from an ordinary American voter, a Make America Great Again, or MAGA, supporter, or a niche cultural community. The apparent domestic voice gives an imported narrative the costume of grassroots dissatisfaction.

Sofia Jadler:

That distinction matters legally, though not in the simple way people sometimes assume. Deception may violate a platform’s rules without automatically violating a federal criminal statute. X’s authenticity rules address manufactured identities, deceptive impersonation, coordinated inauthentic activity, ban evasion, and account repurposing. But a platform violation is not itself proof of fraud, illegal foreign spending, or a violation of the Foreign Agents Registration Act, commonly called FARA.

Red Velhouse:

Let’s take the account that presented itself as Coolture. It reportedly had more than 200,000 followers. Why does that matter if we cannot tell how many saw or believed the new content?

Omar Seidren:

Because audience access is an asset. Building a credible account from zero is slow and uncertain, while acquiring, hijacking, or repurposing an established account can provide initial distribution. But 200,000 followers does not mean 200,000 persuaded voters. Graphika could not determine whether the account was hijacked, sold, purchased, or otherwise transferred, and we lack complete impression data. A large follower count is potential reach, not demonstrated influence.

Ann Tofado:

That distinction is politically essential. Reach can matter even when persuasion is weak, because content may be screenshotted, quoted, or used by genuine partisans to reinforce existing beliefs. Yet researchers have also said Spamouflage has historically been poor at convincingly mimicking real Americans. The operation can be persistent and adaptive while still being bad at winning a real political conversation.

Red Velhouse:

Does that make it mostly noise?

Ann Tofado:

Not necessarily. Noise can be useful when it makes reliable information harder to distinguish from manufactured grievance. But the reporting does not establish changed opinions, altered turnout, shifted candidate support, or compromised voting systems. We should not confuse coordination with electoral effect.

Omar Seidren:

The strongest demonstrated capability is account coordination and persona manipulation. The AI-generated video matters, but it is not the whole story. One account used such a video to argue that the United States is an oligarchy rather than a democracy. That shows synthetic media can package a message effectively; it does not show that AI powered the entire operation or reached a decisive audience. The less glamorous infrastructure—account history, timing, and amplification—is probably more important.

Omar Seidren:

Generative AI still lowers the cost of making profile images, videos, and persona-specific material. But a polished video attached to an empty account is standing in an empty room. The interesting combination is cheap synthetic content plus occasional access to aged or established accounts. That could make the operation more scalable, although the report gives one explicit AI-video example, not a quantified industrial pipeline.

Sofia Jadler:

Scale alone does not answer the legal question. Unpaid organic posts are different from purchased political advertisements, campaign expenditures, or coordination with a candidate. U.S. campaign-finance law broadly bars foreign nationals and foreign principals from certain election-related contributions and disbursements, while issue advocacy not connected to an election can be lawful. We have no evidence here of paid advertising, campaign payments, or candidate coordination.

Red Velhouse:

What about FARA? Could an overseas influence operation simply be required to disclose itself?

Sofia Jadler:

FARA generally requires covered agents of foreign principals engaged in political activities in the United States to disclose their relationship and activities. But we do not know the operators, whether there was a U.S.-based intermediary, or whether the required legal relationship can be established. Registration is a disclosure regime, not a general ban on advocacy. Anonymous overseas operators also create problems of jurisdiction, service, and proof.

Ann Tofado:

That is why policymakers should avoid collapsing every foreign-originated political message into “election interference.” The phrase can cover trying to influence a vote, undermining trust, repressing diaspora communities, or suppressing criticism of Beijing. Spamouflage has been associated with targeting Chinese dissidents and diaspora audiences as well as U.S. political narratives. Those objectives can overlap, but the response may differ for each.

Red Velhouse:

Sofia, does the law have a useful answer when the harm is institutional distrust rather than a specific false statement or expenditure?

Sofia Jadler:

Often there is no clear direct answer—and that is deliberate. Political speech receives strong protection, including anonymous speech, even when it is unpleasant or misleading. Legal hooks become stronger with impersonation intended to defraud, unauthorized access, threats, campaign spending, or a provable agency relationship requiring disclosure. Without those facts, institutions may rely more on platform rules, sanctions, intelligence work, and public attribution than prosecution.

Ann Tofado:

Disclosure can cut both ways. Exposing an inauthentic network may help users discount it, but dramatic attribution can amplify the content, invite partisan arguments about censorship, and encourage genuine political actors to label any damaging criticism foreign-made. The challenge is explaining the evidence without treating every organic disagreement as an operation.

Red Velhouse:

What would count as genuine escalation before or after November third?

Omar Seidren:

I would watch for new clusters coordinating across platforms, candidate-specific targeting, paid promotion, or evidence that posts are reaching authentic users outside the network. A dormant account posting a few times is one thing; a network sustaining replies, reposts, video production, and cross-platform movement is another.

Ann Tofado:

Post-election claims may be especially important. If accounts move from general distrust to claims that a particular result was stolen, or encourage threats and violence, that changes the risk profile. It would show an attempt to convert ambient skepticism into a concrete legitimacy crisis. But we should describe what is observed rather than assume every post will produce mass mobilization.

Sofia Jadler:

Legally, investigators would look for facts connecting speech to conduct: money, direction, coordination, access, threats, or deception causing a legally recognized harm. The operator’s advantage is remaining inside ambiguity—anonymous, unpaid, and politically expressive. The institutional response should preserve evidence and use the narrowest defensible tool, rather than announce a conclusion the facts cannot support.

Red Velhouse:

Before we close, Omar, is this a new method or an evolution of an old one?

Omar Seidren:

It is an evolution. Graphika has monitored Spamouflage since at least 2019. Earlier work described a progression from general criticism of the U.S. political system, to election narratives during the 2022 midterms, to more developed American-looking personas before the 2024 presidential election. The 2026 activity adds account repurposing, selective audience access, closed-loop amplification, and occasional generative media. The tools change, but the engineering problem is still trust.

Ann Tofado:

The political lesson is that resilience cannot mean treating the operation as either omnipotent or harmless. It is not established that these accounts changed an election. It is established that foreign-linked actors are trying to make political identities and grievances harder to authenticate. Institutions must respond without allowing that response to become another source of partisan delegitimization.

Sofia Jadler:

The legal lesson is similarly narrow but important: attribution is not liability. Graphika can assess the broader Spamouflage operation as linked to Chinese state actors while using more limited language about particular accounts. That distinction protects accuracy and preserves room for enforcement if later evidence shows paid election activity, a U.S.-based agent, unlawful access, threats, or another specific offense.

Red Velhouse:

The unresolved issue is impact: these accounts clearly show coordination, deceptive identity-building, and efforts to erode trust, but the public evidence does not yet show that they changed voter behavior or affected an election outcome. Worth watching are new cross-platform clusters, authentic-user reach, paid or candidate-linked activity, post-election claims that voting was stolen, and evidence connecting anonymous accounts to identifiable operators or intermediaries. Sources and references for this discussion are available with the episode at Factolio.com.

---

Sources and References

These sources supported the factual material used in this discussion. Factolio’s panel discussion is AI-generated from researched evidence and is written in original language.

1. Graphika — The #Americans – Season 2: Spamouflage and Inauthentic Pro-China Accounts Target the 2026 US Midterms (PRIMARY)
2. Associated Press — Graphika finds China-linked Spamouflage accounts posing as US voters on X (NEWS)
3. Graphika — The #Americans (PRIMARY)
4. Office of the Director of National Intelligence — Election Security Update, October 21, 2024 (PRIMARY)
5. FBI and Cybersecurity and Infrastructure Security Agency — Just So You Know: Foreign Threat Actors Likely to Use a Variety of Tactics to Develop and Spread Disinformation During 2024 U.S. General Election Cycle (PRIMARY)
6. X Help Center — Authenticity (PRIMARY)
7. U.S. Department of Justice — Foreign Agents Registration Act (PRIMARY)
8. U.S. Department of Justice — Foreign Agents Registration Act Frequently Asked Questions (PRIMARY)
9. Federal Election Commission — Foreign nationals (PRIMARY)
10. Legal Information Institute, Cornell Law School — 52 U.S. Code § 30121 – Contributions and donations by foreign nationals (OTHER)