AI Is Multiplying What One Person Can Do. HOWEVER It Is Not Multiplying Their Judgment. And That's the Problem.
CrowdStrike says a suspected individual attacker used AI-assisted tooling in a campaign targeting at least nine South Korean banks. The investigation is still developing. Whatever the final attribution, the ratio is what matters: one operator, organizational-scale reach.
For most of history, creating organizational-scale consequences required an organization. It demanded people, expertise, coordination, time, and review. AI is beginning to weaken that relationship, allowing individuals to exert leverage previously reserved for institutions.
Researchers already use "uplift" to describe how AI increases speed, scale, or sophistication. I use "asymmetric uplift" for the more consequential case: when a relatively small increase in human-plus-AI capability produces a disproportionately large increase in possible consequence. The mechanism itself is neutral. A founder with no analyst team can suddenly perform research that once required five people. A local business can produce sophisticated campaigns without an agency. A scientist can interrogate enormous literatures. A teacher can personalize material for hundreds of students. The same multiplier that democratizes capability also democratizes consequence.
The core problem is that capability uplift is not judgment uplift. AI can make you faster without making you wiser. It can make you more persuasive without making you more correct. It can let you act at greater scale without improving the quality of the decision being scaled. Capability can compound faster than judgment. Or, more precisely: AI can multiply the consequences of a decision without improving the decision that produced them.
Consider a simple thought experiment. A person makes a mediocre decision manually. Consequence: maybe it affects 10 people. The same mediocre decision, made with agents, automation, APIs, distribution, and autonomous execution, could affect 100,000 people. Nothing about the original judgment improved. Only its reach did. The dangerous variable may not be how intelligent the person becomes. It may be how far their judgment can travel.
This gives us a better interpretation of AI's power. People keep asking: how much smarter will AI make us? Perhaps that's the wrong question. A more pertinent one is: how much more consequence can one person create? This applies across cyberattacks, finance, media, persuasion, commerce, science, software, and operations.
There's a second-order issue here. Organizations were not only collections of talent. They were also friction. Multiple people meant review, debate, permission, specialization, delay, and accountability. Some of that friction was annoying. Some of it was protective. AI can remove those bottlenecks. That is good when the judgment is good. It is much less good when it isn't. The same friction AI removes in the name of productivity may also have been part of the governance system.
This is the problem Recursum is designed around: not merely increasing how much one person can do, but improving how well that person can reason before capability becomes consequence. Capability uplift without judgment uplift is leverage without steering. The more consequence an individual can create, the more important contradiction, verification, provenance, authority boundaries, and consequence tracking become.
The promise of AI is that one person may soon accomplish what once required an organization.
That could be extraordinary.
It could also expose something we have underestimated.
Organizations did not merely provide capability. They also distributed review, responsibility, disagreement, and accountability.
AI can compress the capability without automatically recreating those safeguards.
What happens when one person acquires organizational-scale power while retaining individual-scale judgment?
That may be one of the defining governance questions of the AI era.
AI is multiplying what one person can do.
Now we have to make sure judgment can keep up.
TL;DR: AI is enabling individuals to achieve organizational-scale operational power, as evidenced by a suspected AI-assisted attack on nine banks by a single operator. This "asymmetric uplift" means a small increase in human-plus-AI capability can yield disproportionately large consequences. While AI makes individuals faster and more capable, it doesn't necessarily improve their judgment or accountability.
This creates a dangerous imbalance where the reach of a decision far outstrips the quality of the judgment that produced it. Organizations historically provided friction, review, and accountability. AI removes these bottlenecks, making it crucial for individual judgment to keep pace with amplified capability.
By Ernesto Verdugo, AI Architect, Founder of Verdugo Labs, and Creator of Recursum. He builds systems for what happens when human and artificial intelligence stop working separately.
Links: