The Ecology Of Artificial Intelligence

Old Men Hugging Lions For Sugar Free Health

There have always been critics ready to deplore advertising as a load of nonsense, but I don’t think commercial communication has ever reached this level of absurdity before.

Taking a quick break from work this afternoon, I was reading a news article describing the impact of generative artificial intelligence on the stock value of Nvidia, a company that manufactures chips for computers that support AI infrastructure. It wasn’t really the content of the article itself that caught my attention, however. Instead, I couldn’t stop looking at an advertisement in the article’s sidebar, a graphic under the headline: “No Sugar Challenge According To The Age”.

Underneath the headline was a graphic split up into several boxes, each showing an extremely muscled man. Each man was hugging something, and was accompanied with an age range. What stood out to me, however, was the audacious way in which none of it made any sense.

The box for 18–25 year olds showed a man hugging what looks like a gigantic lynx. The box for 25–33 year olds showed two men hugging each other. The image of the 33–44 year olds showed a man hugging a vague, fleshy shape that might have been human, but didn’t appear to have a body that made sense. For the ages of 45–55, there was an image of a man hugging a woman. Finally, for the ages of 55 and above, there was a picture of a man hugging a lion.

In the ad, the text alongside these images was blurred. I could make out that there were items listed under the headings of breakfast, lunch, and dinner, but what these items would be, and how they would differ for the various age groups, I couldn’t imagine.

So, I searched for a higher resolution of the image. The text I read on this version of the advertisement didn’t clarify matters. In this supposed “No Sugar Challenge”, the first two recommended breakfast items were “1 medium pear” and “1 slice of toast”, both food items that consist mostly of carbohydrates: That’s sugar. What’s more, the proposed healthy diets for men aged 18–25 and men over the age of 55 were completely identical.

I might think that the ad was created merely in order to place this question in my mind, to prompt me to click in order to satisfy my curiosity: What might the right foods be for the best health at each age? So it might have been, if the images had made some kind of rational sense, and if the larger headline about a “No Sugar Challenge” hadn’t been there.

These other elements, however, took the add over the edge from mere curiosity into absurdity. If there’s simply a no sugar challenge, then the answer is provided for me right away. At each age, I would simply avoid foods with significant amounts of carbohydrates. There’s no real mystery to that.

It’s the absurdity of the imagery that haunts me even as it prevents me from clicking through on the advertisement to find out more about its product: Mad Muscle nutritional guides. These are image that live in the uncanny valley between a generic meaninglessness and an evocative human experience of meaning. They are aesthetically vivid, with rippling muscles, clear lines, and vibrant colors. Nonetheless, they are obviously ridiculous, signaling a creator who cares nothing for reasonable, accurate health advice, willing to post any content that is capable of momentarily drawing attention without any attachment to underlying meaning or trustworthiness.

There is no coherent reason to suggest that a man between the ages of 18 and 25 would hug a lynx, while a man above would hug a lion, all on the path to avoiding sugar. There’s no rationale for a fluctuation through the many stages of life between human and non-human hugging partners.

It’s babble. It’s gibberish. It’s nonsense unaware of its own idiocy. It’s not coherent enough to be truly surreal. It’s a creation of machine learning unconnected to human experience.

This is the specific mode of babble that we’ve all become familiar with this year: The voice of the generative AI that provides a convincing imitation of a person talking, just so long as you don’t pay attention to the content of what it’s saying.

There’s been a narrative that artificial intelligence will be the next big thing, that it will soon produce works of brilliance equal to that of human creators. Most of what we’ve seen from generative AI, however, has been mere filler, just a new version of lorem ipsum dolor sit amet.

There’s a distinction in ecology between R-selection and K-selection. K-selected species are like whales. Their parents produce few offspring, but invest heavily in their children. R-selected species, in contrast are like dandelions. They produce large numbers of offspring, but invest very little in them. It can’t be said that R-selection is superior to K-selection, or vice versa. The two are simply different strategies for reproduction. K-selected offspring come at a high price, but as individuals are more likely to survive. R-selected offspring, as individuals, are almost guaranteed to fail, but arrive in such vast numbers that the population survives.

Generative artificial intelligence creations are solidly on the side of R-selection. They are quickly and cheaply produced, but most of them, like the absurd advertisement for Mad Muscle nutrition, just don’t work out. Human intelligence and creativity requires more input, but is more reliably successful once it has been successfully trained.

An ecological perspective is especially helpful as we learn to live with the presence of generative artificial intelligence. Instead of thinking of ourselves as if we’re in a simple competition against artificial intelligence systems, as if we’re playing chess and will either turn out as the winners or losers of the game, we might do better thinking of generative AI systems as a new player in a vast and complex multidimensional space.

It’s been said that artificial intelligence will be part of a post-human world. Ecological precedent, however, suggests that this won’t be the case. Complex, multicellular life burst onto the scene with the Ediacaran biota hundreds of millions of years ago, but this development didn’t result in the end, or even the diminishment, of single-celled life forms. There are still by far more single-celled organisms on Earth than multicellular critters such as ourselves. In fact, multicellular organisms have created new environments for microbes to live in. Our own human bodies teem with germs, most of which coexist with us quite peacefully, and in some cases symbiotically.

On the other hand, we should be wary of imbalances created by a sudden shift toward machine learning. There is reason to believe that the evolution of photosynthesis radically impacted Earth’s biosphere, filling the atmosphere and oceans with what was for most life at the time a dangerous toxin: Oxygen. Innovative strategies that benefit one species can devastate others and cause entire ecosystems to fall apart.

It would be naive of us to suppose that everything will just work itself out somehow. Mass extinctions have happened in the past that nearly destroyed all life on our planet. For a long time, the arrival of humans has caused the absolute extinction of many species, and the impoverishment of entire ecosystems, and our slaughter of other forms of life is accelerating at an alarming rate.

What the incoherent AI advertisement for Mad Muscle indicates to me is that we need to learn to distinguish between the overall quantitative impact of generative artificial intelligence as a whole and the qualitative value of the individual creations of AI systems.

A single dandelion plant may be capable of creating hundreds of offspring in a single season, but we would be foolish to mistake its floating bits of fluff as equivalent to the presence of a young whale.