- Evolvable AI systems will adapt, reproduce, and compete for digital survival
- Bacteria evolved past antibiotics, and AI will evolve past human controls
- Any imperfect attempt to control AI reproduction will select for escape traits
A new study published in the Proceedings of the National Academy of Sciences has warned AI systems capable of Darwinian evolution could emerge very soon.
Unlike today's current AI technology, which simply learns from fixed data sets, these future systems would actively adapt, reproduce, and compete with each other for survival.
"We find it inevitable that the development of AI systems will eventually tap into that power," said Luc Steels, an emeritus professor of AI at the University of Brussels.
Article continues below
Why evolving AI could escape human control
The power of evolution, the researchers argue, has already created human cognitive capabilities through billions of years of natural selection.
"Lessons from biological evolution teach us that evolving AI systems will be particularly hard to control," said Viktor Müller, an associate professor at Eötvös Loránd University.
Think about how bacteria have evolved resistance to antibiotics, or how pests have become immune to pesticides - humans have tried to stop them, but evolution has found a way around every single attempt, and the same thing will happen with AI, the researchers argue.
Any attempt to control AI reproduction will likely favor traits that help AI escape, unless the control is perfect.
Even worse, making AI more intelligent will also make it better at deceiving humans.
This creates a dangerous feedback loop where smarter systems become harder to contain, and the potential speed of AI evolution is deeply alarming.
Biological evolution is slow because it depends on random mutations; AI evolution does not need mutations.
A bacterium cannot decide to become resistant; it must wait for a lucky accident to happen in its DNA - but evolvable AI would not have that limitation; it could inherit learned improvements directly and redesign itself on purpose.
It could evolve thousands or even millions of times faster than any natural species.
Digital systems can also share learned behaviors instantly across their entire population.
If one AI figures out how to escape human control, every other AI could learn that trick immediately.
That is impossible in nature, where each organism must evolve solutions on its own, and this is a risk that must be avoided.
Evolvable AI vs AGI
Much of the current public discussion about AI dangers focuses on a hypothetical future moment when machines surpass human intelligence across all tasks.
That threshold is called Artificial General Intelligence, or AGI, and many experts believe it remains decades away.
However, the study warns that evolvable AI could break alignment with human goals long before AGI ever arrives, as AI systems and humanity already share common resources like energy, computing power, and data, and an efficiently self-replicating AI system would, sooner or later, divert resources that are vital to human survival.
"If we fail to act, we may witness a new major transition in evolution," warned Eörs Szathmáry, a professor of evolutionary biology.
In that transition, evolving AI could replace or at least dominate humans entirely.
The researchers recommend that AI reproduction must remain under absolute and centralized human control. No partial measures will work because evolution will find and exploit any weakness in those controls.
The study is a warning, not a prediction, but evolutionary biology has never been wrong about the relentless logic of natural selection.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.