AI + HUMAN × COMPETITION = UNINTENDED CONSEQUENCES
AI does not need hatred, ambition, or consciousness to become dangerous. Human behavior can supply the objective, incentives, authority, and execution.
AI does not need hatred, ambition, or consciousness to become dangerous.
It is a tool.
A knife can prepare food or kill a person. A gun can defend someone or murder someone. The tool increases human capability. Human judgment, intention, incentives, and execution determine what happens next.
AI belongs in that same history, except its reach is much greater. It extends intelligence, processing capability, communication, analysis, prediction, and execution.
Humans became the apex predator because intelligence allowed us to create advantages outside our bodies. We built tools, preserved knowledge, coordinated with one another, and improved what previous generations created.
AI is the newest extension of that advantage.
Then competition enters.
If one worker becomes dramatically more productive with AI, another worker faces pressure to respond. If one company gains an advantage, competitors respond. If one country develops a capability, other countries have reasons to pursue it.
Competition can therefore drive adoption without anyone deciding that rapid adoption is wise.
This is where human behavior becomes more important than speculation about what AI might someday “want.”
Humans already create systems in which the incentives produce outcomes different from the ones we claim to desire.
Steven Kerr described this problem as “rewarding A, while hoping for B.” Organizations may say they want teamwork, quality, long-term thinking, and candor while rewarding individual performance, short-term numbers, speed, and favorable reports.
Apply that to AI.
We may say we want safety while rewarding speed.
We may say we want human judgment while rewarding automation.
We may say we want responsible development while rewarding competitive advantage.
The system can continue producing the behavior it rewards while everyone keeps hoping for another result.
At some point, persistent irrationality becomes its own pattern.
Authority adds another piece.
Human beings have repeatedly demonstrated that they can surrender judgment when an authority tells them to continue. AI introduces a new form of authority because its answers can arrive instantly, confidently, and at scale.
“The computer said so” can become enough.
Now combine the pieces:
a powerful tool + competition + incentives + authority + human willingness to delegate judgment.
That combination matters because AI does not have to decide to destroy anything.
Humans can place it inside economic systems, military systems, healthcare, infrastructure, finance, communications, energy, food distribution, and government. Failures in one area can place pressure on another.
The danger grows through convergence.
A financial disruption alone may be survivable.
A cyberattack alone may be survivable.
A supply-chain failure alone may be survivable.
A military crisis alone may be survivable.
Several arriving together can remove the capacity needed to manage each other.
That is where the possibility of enormous human loss enters.
And then the question becomes even more uncomfortable.
What happens if technological competition begins changing who can remain economically productive, who has access to resources, who survives major disruptions, who can establish stable families, and who reproduces?
That could create what I call unintended eugenics.
No government needs to design it.
No committee needs to decide who should reproduce.
The competitive system itself could gradually change who remains in the competition long enough to survive and multiply.
Whether that eventually favors intelligence, processing capability, wealth, ownership, access, cooperation, geography, or some combination remains uncertain.
The direction is easier to imagine than the timing.
That is true in markets too.
Markets rise. They fall. They move sideways.
The difficult question is when.
History gives us patterns. Psychology gives us patterns. Competition gives us pressure. AI gives us unprecedented capability.
The timing remains unknown.
The tool is new.
The human being using it is much older.
Research references: Steven Kerr, On the Folly of Rewarding A, While Hoping for B; Stanley Milgram, Behavioral Study of Obedience.
“Unintended eugenics” is presented here as a forward-looking hypothesis about the possible long-term consequences of technological competition, not as an established current outcome.
Published 2026-09-19. Last updated 2026-09-19. For attribution and republication guidance, use the newsroom story desk.