Skip to main content

The Information of Intelligence

Defining intelligence anywhere in the universe is tricky, but there’s one thing that might provide a clue

Credit:

C. Scharf 2018 with thanks to W. Shakespeare, Henry VIII; Credit: NASA, ESA, Hubble and F. Ferraro (background image)

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


When we talk about life in the universe, or intelligence in the universe, we inevitably run into a bit of a problem. That’s because we (meaning scientists) don’t really have a very good definition of either phenomenon. The ‘what is life?’ question is a perennial one, and I think part of the issue is that living systems embody multiple phenomena in different mixes depending on the organism in question. 

Another part of the issue is that we don’t have a first principles theory of life – we don’t have a mathematical, predictive framework. In part that’s because we don’t know whether life is substrate independent; capable of utilizing a variety of building blocks, from organic molecules to inorganic molecules, to software. 

Intelligence seems to fall into a similar hole. I recently participated in a very informal discussion in a room full of scholars – from neuroscientists to AI experts and biologists – where an effort was made to define intelligence and suggest tests of intelligence. My rather feeble contribution was to say that I could only add large question-marks to the list. But others were bolder and we got ideas that ranged from intelligence being about the ability to project into the distant future to intelligence being the ability to deceive.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


Much like the debate over a definition of life, I felt that each of these options were useful, but probably only in specific cases. And that got me thinking about whether there was a property of intelligence that was much more easily quantifiable – something that could even form a sliding scale, a true measure of minds. Not in the sense of an intelligence test, with all of its attendant pitfalls and biases, but a more general evaluation of a property seen across a Darwinian landscape of variation and adaptation. 

The working hypothesis that I came up with (and blurted out to some rather blank stares in the room) is that the propagation of information into the future might offer a clue. 

How might this help? Well, humans do something rather strange and interesting. As a species (as well as individuals) we generate persistent information that exists outside of our genome. 

For example, Shakespeare wrote 37 plays with a total of 835,997 words (yes, really, the data is here). Those words, in that precise sequence, were not a part of his heritable DNA, yet here they are centuries later. To me that suggests a very particular phenomenon is at work.

Something as different as a termite colony might construct a hugely intricate nest of tunnels and structures that ends up being preserved as a rocky structure for tens of millions of years. But the final information content is not the same as Shakespeare’s plays. (There’s probably a nifty computational research project there; to evaluate information entropy in these radically different edifices).

And today humans crank out ridiculous quantities of information-rich stuff. Estimates put the amount of device-based data production at around 2.5 quintillion bytes a day (2.5x1018 bytes per day).

In very crude terms, even with 7.5 billion living humans, each with about 3.2 billion base-pairs in their DNA (and ignoring epigenetic factors, among other complications), our biological information content pales by comparison to our extant non-biological information.

To my mind that is an interesting, measurable consequence of this thing that we call intelligence: The ratio of genomic information content to external, persistent information baggage. Our very first oral history and our first cave painting initiated an information-rich future that was at the extreme end of what other organisms on Earth are capable of. 

I don’t think this magically solves the question of what the best measure of intelligence is (across species). But it might be useful as we try to gauge the intelligence of the new machine species we’re busy molding out of silicon. It might also be useful as we sniff around the cosmos looking for evidence of other life and other minds.

An alien mega-structure (should such a thing actually exist outside of our feverish imaginations) might just be a larger version of a termite mound, but it could also betray the properties of the minds behind it by its information content.