When I “Asked Nature” About Artificial Intelligence, Here’s What I Learned
“The only thing necessary for the triumph of evil is for good men to do nothing.” — Edmund Burke
When I first heard this quote, it was in the context of 18th century European history and the French Revolution. I was surprised to discover how applicable that same idea is today! Read on and I’ll explain what I mean.
In my first blog in Asking Nature back in July, I shared the journey I took with 30 Days of Reconnection. In September, I shared How I Found the Circular Economy: Biomimicry and the Power of Design. Since then, I’ve been thinking of ways to combine my curiosity about artificial intelligence (AI) with the sense of awe I feel when I reflect on designs from nature. I was encouraged to find examples of designs on AskNature.org.
As we think about biomimicry and developing machine intelligence, we have billions of years of natural engineering we can learn from to enhance our AI design. Here are a few surprising examples from Ask Nature and other sources where AI designers can tap into nature’s blueprints.
Example 1: Energy Efficient Cat Brains
Compared to a cat’s brain, computers are far from efficient. The majority of current computer signaling is still linear (think a bridge between two islands — point A to point B). But cat brains are different: pound for pound and watt for watt, they are extremely more efficient than our best computers. To catch up with the cats, researchers are developing ways to replace transistors with synapse-inspired devices called memristors. But that’s just one part of the puzzle.
An article in WIRED describes how one researcher, Geoffrey E. Hinton, studied those feline brains and made a tremendous impact on the field with his findings. Hinton and his team figured out that a cat’s brain processes sensory information in stages, one layer at a time, in particular parts of the brain.
To duplicate this process in computing, his team started writing their algorithms to work in a similar way, one layer at a time, modeled on networks of neurons. This approach turned out to be a more efficient way to recognize speech and images, and it is still the foundation of how many of the big tech companies in Silicon Valley conduct deep learning today. To my surprise, some of the best Facebook, Amazon, and Google AI features found their inspiration from the brains of cats!
Example 2: Brainless Slime Molds
A second example of an AI system designers can learn from is something you might not have even thought existed: the brainless slime mold. Don’t underestimate the potential of a slime mold: though they might look like a splatter of mustard, they have an incredibly complex system worth learning from. The special function they perform is memory retention and communication, even though it has no “control center” (aka, a brain). We don’t know for sure what the mechanism is, but scientists think it might be related to the transfer of genetic information between slimes. Others believe it may be chemical communication. Whatever it is, the key takeaway is that AI systems might be made more efficient by experimenting with new designs that don’t require a traditional control center. We can also tap into the wisdom of the slime molds’ mutual cooperation and swarm intelligence techniques. Whether metaphorical or literal, the translations into design learned from this biological champion shows that there are ideas in nature that can surprise and inspire us. Who would have thought something could operate without a brain unless they saw nature do it first?
Example 3: Spider Webs — An Extension of Mind and Memory?
Even though it has eight eyes, The Orb Weaver Spider can only sense a little bit of light, dark, and movement. But despite this setback, the silk in their webs is five times stronger than steel. And, what’s even more interesting and mysterious is the way the web and the spider interact. The web can be tightened to improve prey detection. When the web moves, the spider feels the movement and gets a sense of its prey’s location. The spider remembers where in the web they store their wrapped-up prey even though its brain is only the size of a pinhead.
Scientists have a theory that this spider’s mind is actually connected to the strands of its web, forming a map of the spider’s memories. The way the Orb Weaver spider uses its web to sense its environment and expand its memory could be an interesting design inspiration for new kinds of AI — imagine a new kind of computer intelligence that could operate with the efficiency of these creatures! We’ve already learned the tensile strength from engineering materials like the web itself, and we continue to learn new strategies from this incredible organism. (Bonus fact: did you know an Orb Weaver spider’s web can also contract and relax in response to humidity?)
Responding to a changing environment is crucial in an ecosystem and in our AI systems. What else might we learn from the more than 45,000 known spiders alive today?
The Overall Picture
No matter what kinds of inspiration we find in nature, and no matter how optimistic AI might make us feel about the future, we should be cautious. When plastics were first invented, everyone thought they were a great idea… but most people never imagined the detrimental environmental effects. We need to consider the full effect of these brilliant ideas before they get out of control with unintended negative side effects. If we don’t think carefully about the future, we could end up with an AI version of the Pacific Garbage Patch! Or worse… the Matrix.
Some are already thinking about the need to consider how ‘blind technology without compassion is ruthless’, and we must take these considerations seriously. AEye, an autonomous perception pioneer who is creating technology for autonomous vehicles, has looked to the human visual cortex for enhancing its iDAR™ system, a new form of intelligent data collection that seeks to think like a robot, but perceive like a human. If humans have the gift of looking ahead to the future, we can take all these warnings into consideration and look to nature as our mentor to not only fit in with the existing environment, but to give back to it generously.
To prepare for the future we need diverse teams that can look at AI from many different points of view. Everyone should get involved in shaping the future of AI — don’t leave AI to the “experts”! We need to bring our passion for the environment and compassion for all of life into the development of AI, before it’s too late. As exciting as AI is, it is important to look at the potential of AI through a skeptical lens. We must never forget that every breakthrough technology often has a dark side.
After looking at these design ideas, I’ve reached this conclusion: a natural world without AI would still be a wonderful place — but living in a world of AI without nature is a boundary we should never cross.
Sophia Stiles is a high school student based in California. In her free time she enjoys drawing and taking walks with her family.