Let's take the squid.

He has more neurons than a mouse, which in turn are not concentrated in one brain, but are also distributed over his eight arms, which can act quite independently as a result.

Squids can navigate mazes, tell people apart, open jars - they are the geniuses of mollusks, so to speak.

At least nothing comparable is known about snails or clams.

But how intelligent is an octopus actually?

Alexander Armbruster

Responsible editor for business online.

  • Follow I follow

Paul Thagard is a philosopher and cognitive scientist, a member of the Royal Society of Canada and a former professor at the University of Waterloo.

He has now written a lively volume with the aim of classifying and making tangible what intelligence is all about.

This is expressly against the background that, in addition to humans, not only animals have these skills, but also computers are increasingly gaining abilities that we characterize as intelligent.

Our brain is no longer the measure of all things

The title "Bots and Beasts - What Makes Machines, Animals, and People Smart?" hints at the original approach Thagard is pursuing.

He openly admits that to this day there is no definition of intelligence that experts have agreed on.

Incidentally, this also applies to the aspect of the “artificial”.

Therefore, the author approaches with examples, using six "bots" to explain what AI systems such as Alexa, IBM Watson, Google's translation service, the Netflix recommendation algorithm, the AlphaZero program or self-driving cars are and can do.

He also discusses the extent to which ravens, dogs, dolphins, chimpanzees, bees and squid behave intelligently.

One thing becomes clear: defining intelligence as what an IQ test measures is of little use.

With these examples, Thagard illustrates an increasingly important perspective, especially in the discussion about artificial intelligence: It is about the question of how researchers should proceed in the future to further develop AI systems: As much as the human brain remains a central reference, it is inclusive of the human one Sensor technology is no longer the measure of all things.

Bats use sonar, some bird and fish species can perceive electromagnetic signals, some autonomously driving cars (others do not exist yet) have lidar or GPS and can also locate themselves and analyze their surroundings in other ways than humans are able to do.

The subject thrives on the exchange of different disciplines

Far-reaching questions also arise here for practice, because the AI ​​systems that are currently so popular, which are described as artificial neural networks or "deep learning", increasingly recognize colors, shapes, sounds and therefore whole with the help of high-resolution cameras or corresponding audio hardware general patterns that humans cannot easily see or hear.

This is a great opportunity when it comes to checking components for safety-relevant machines, simulating new molecular constellations for medicines or detecting common behavior.

However, it is also a source of danger because the users of such systems have to ensure that they do not make mistakes themselves.

Meanwhile, Thagard continues to ask.

Based on his bots and beast examples, he discusses the relationship between consciousness and intelligence, empathy and emotionality.

In addition, he makes policy recommendations and recalls two important findings in AI research from recent years: The subject thrives on constant interdisciplinary exchange between computer scientists, neuroscientists, biologists and other specialists.

And humans, no matter how much the AI ​​pioneers focused on their skills after the Second World War, are in many respects not the sole yardstick when it comes to the full potential that computers may one day develop.

Paul Thagard: "Bots and Beasts".

What Makes Machines, Animals, and People Smart?

The MIT Press,Cambridge 2021. 312 pp., hardcover, €27.