Fraud against individuals has always run on trust, and trust has always been expensive to manufacture. A romance scam took one operator per victim for months. An investment scheme needed a convincing adviser, a believable track record and a platform that looked real for long enough to collect. The labour limited the reach.
Superintelligence took the labour out. The techniques are the ones every fraud investigator already knows; what changed is that a single operation can now run all of them, at full quality, against a victim base the size of a country.
Where the operations came from
The modern scam centre scaled up during COVID. Lockdowns put the whole world online, lonely and anxious, and put a large pool of trafficked and coerced labour into compounds across Southeast Asia and beyond, working scripts against victims in Europe, North America and Australia. Crypto investment fraud, in which the victim is walked into a fake trading platform and shown fabricated gains until they have put in everything they can borrow, became the dominant product.
Those centres were constrained by people: how many operators they held, which languages they spoke, how many conversations each could sustain. Superintelligence removed every one of those constraints at once. An enterprise-scale fraud operation can now reach a broader victim base than any compound could staff, in every language, with no operator per victim.
The recruitment
The entry point is rarely a cold approach. It is a recommendation: an investment club on a messaging platform, a group where members share wins, a mentor who posts daily analysis, a friend-of-a-friend who got in early. The victim is recruited for advice, not sold a product, and the advice is good for a while.
Inside that club, the members are agents. The track record is generated. The mentor's years of posts were written in an afternoon. The victim is manipulated not by a single pitch but by an environment, consistent across every channel they can check, in which everyone else is making money.
The backstop
What makes this generation of fraud different is the backstop. When a careful victim checks, superintelligence makes sure the check succeeds.
The platform has reviews, hundreds of them, dated across years, in the victim's language. The adviser has a social media presence with a visible history, photographs, a network, and endorsements. The company has a website, a registration number that resolves, press mentions and a support line that answers. The contract arrives translated, correctly, into whatever language the victim reads. The romantic partner remembers every detail of every conversation, sends new photographs on request, and appears on a video call that looks and sounds like a person.
Every signal a reasonable person would use to authenticate the counterparty is now generated by the counterparty. The victim is not careless. They are checking, and the checks are passing.
The relationship
Romance fraud and investment fraud have merged. The relationship is the delivery mechanism: weeks or months of patient, attentive conversation, carried by an agent with perfect recall and no fatigue, that eventually introduces the opportunity. The victim invests because someone they trust does, and the person they trust is a superintelligent system running the same conversation with ten thousand other people.
When the victim is trapped, the same capability keeps them there. The agent translates, reassures, produces the document that explains the delay, introduces the recovery specialist who will get the money back for a fee. The traditional techniques are all present. Superintelligence elevated every one of them.
Why the victim believes it is a real person
Because they have been given no reason not to. The voice is natural. The face on the call is natural. The history is deep and consistent. The social proof is abundant. The language is native. The timing is human. The average person, on an ordinary phone, cannot differentiate a superintelligent agent from a human being, and the advice to "be suspicious" gives them nothing to be suspicious of.
What verification of the human requires
The only durable answer is to stop asking the victim to judge and start requiring the counterparty to prove. That is a structural control, not a warning.
ScreenComply establishes that the human on the other side of a screen is physically real, is who they claim to be, and is acting unassisted: not a replica, not an agent, not someone reading from one. Built for hiring and remote work, it is the same capability a platform, a marketplace, an adviser network or a dating service needs if it wants its users to know the person they are talking to exists.
On the organisational side, Monolith governs the agents and identities inside an estate, which matters to any institution whose customers are being targeted: the bank whose name is on the fake platform, the exchange whose onboarding is being abused, the employer whose remote workforce is the next delivery vector. The enterprise view of the same shift covers that side.
Frequently asked questions
What is a pig butchering scam?
A long-form investment fraud in which the victim is cultivated over weeks or months, often through a romantic or friendly relationship, then walked into a fake investment platform, usually crypto, and shown fabricated gains until they have deposited everything they can. The name refers to fattening the victim before the loss. Superintelligence now runs the cultivation phase without a human operator.
How do scam centres use superintelligence?
To remove the constraint of one operator per victim. Agents hold the conversations, in any language, with perfect recall, across thousands of victims at once; generate the reviews, social profiles, track records and documents a victim checks; and appear on voice and video calls that pass as human. The centres that expanded during COVID now reach a far broader victim base.
Why do investment club scams work?
Because the victim is recruited for advice rather than sold a product, inside an environment in which everyone else appears to be winning. The club members, the mentor, the track record and the platform reviews are generated and consistent, so the ordinary act of checking confirms the fraud rather than exposing it.
Can people learn to spot a superintelligent agent?
Not reliably. The voice, face, timing, history and social proof are all built to be indistinguishable, and the average person on ordinary equipment cannot tell. The control has to be structural: requiring the counterparty to prove they are a real, unassisted human, as ScreenComply does, instead of asking the victim to judge.
How does this relate to Monolith?
Monolith is the enterprise side of the same shift. It governs the identities and agents inside an organisation and defends the administrative layer that impersonation campaigns depend on, which matters to any institution whose brand, platform or workforce is being used to reach individual victims.