Counterfeit Consensus: Why the FTC's Fake-Account Op-Ed Is a Forensics Problem

On October 2, 2026, Sander Gerber — founder and CEO of Hudson Bay Capital and a member of the President's Intelligence Advisory Board — and Eugene Kontorovich, a law professor at George Mason University's Antonin Scalia Law School, published an op-ed in The Washington Post titled "How the FTC and Congress can stop fake foreign social media accounts."

The argument is simple and, frankly, overdue: the Federal Trade Commission already has the statutory authority to police "unfair or deceptive acts or practices." A bot account posing as a human, or a foreign operative posing as an ordinary American — their example is a "Pittsburgh housewife" — isn't an opinion the First Amendment protects. It's a fabricated identity designed to deceive. The authors put it bluntly:

This is a smart legal move. Disinformation fights usually collapse into First Amendment arguments about protected speech and viewpoint discrimination — a fight regulators tend to lose, and should lose. But impersonation isn't speech. It's fraud. Reframing fake accounts as a consumer protection issue — the same category as fake reviews, counterfeit goods, or deceptive advertising — sidesteps the censorship debate entirely and gives the FTC a path that doesn't touch anyone's actual opinions.

The "counterfeit neighbor" problem

The op-ed's most useful contribution isn't the legal theory — it's the mechanism. People calibrate their own views based on perceived social consensus. If your feed looks like "everyone" agrees with a position, you adjust toward it, even subconsciously, even if "everyone" is fifty accounts run by three people in a server farm.

We'd call this counterfeit consensus. It's not that one fake account changes anyone's mind. It's that enough fake accounts, timed and networked correctly, manufacture the appearance of a groundswell — which then gets picked up by real engagement, real journalists, real research notes, and in financial markets, sometimes priced directly into a security.

The op-ed cites two live examples already playing out in 2026:

• OpenAI disclosed a banned influence network — assessed as most likely Chinese state-linked — posing as ordinary Americans in the debate over AI data centers.
• X (formerly Twitter) purged roughly 42,000 accounts using AI to impersonate humans at scale.

Both are textbook counterfeit-consensus operations. Both were only caught because the platforms themselves eventually looked — which is precisely the authors' point about incentives.

The proposed remedy — and its limit

Gerber and Kontorovich propose that the FTC, or Congress, require major platforms to offer users a toggle: an opt-in filter to view only content from verified U.S. residents. It wouldn't be the default. It wouldn't remove foreign or bot content from the platform. It would simply let a consumer choose a feed without manufactured amplification, the same way a label lets a consumer choose products without certain ingredients.

It's a clean, narrowly tailored proposal — and it depends entirely on something the op-ed only gestures at: someone has to actually know which accounts are fake.

This is the part we think about every day. Platforms have the raw signal, but they're graded on engagement, not authenticity — so self-policing has a structural ceiling. If the FTC takes the authors' advice and starts "investigating the scope of the problem," that investigation needs an evidentiary basis that doesn't come from the platform being investigated.

Where forensic detection fits

This is the gap our BADS methodology (Bot Activity Detection System) was built to close — not flagging suspicious accounts on vibes, but cross-referencing account behavior, network topology, and activation timing to prove coordination with evidence that holds up outside a screenshot.

Three things any serious enforcement or litigation effort will need, and that forensic bot-network analysis already produces:

• Network topology mapping — showing which accounts are structurally connected (shared infrastructure, coordinated posting windows, templated language) rather than independently arriving at the same opinion.
• Activation timelines — when a cluster of accounts spun up, went dormant, or activated in lockstep around a specific narrative or event.
• Calibrated confidence, not binary labels — a defensible finding states what the evidence shows and what it doesn't, because "bot or not" is rarely a clean yes/no at the account level; it's a pattern at the network level.

Whether this particular proposal becomes FTC rulemaking, a bill in Congress, or nothing at all, the underlying diagnosis is correct: platforms will not meaningfully self-police counterfeit consensus, because the business model rewards volume over authenticity. Anyone exposed to that gap — brands, funds, legal teams, public figures — needs an independent, evidence-grade answer to "is this real, or is this manufactured?" That's the work.