Korteling2021HumanArtificial

Hans Korteling et al., "Human- versus Artificial Intelligence"

Bibliographic info

⇒ Korteling J E(H), van de Boer-Visschedijk GC, Blankendaal RAM, Boonekamp RC and Eikelboom AR (2021) Human- versus Artificial Intelligence. Front. Artif. Intell. 4:622364. doi: 10.3389/frai.2021.622364

Commentary

⇒ This text provides an alternative view on human intelligence which differs from the anthropocentric view. The anthropocentric perspective treats human intelligence as general intelligence, which then becomes the benchmark for AI intelligence. However, this text suggests we aren't as smart as we think. Consequently, it proposes an approach where we separate human intelligence from general intelligence and allow human and AI intelligence to both have their own strengths and weaknesses. As a result, we shouldn't base AI intelligence solely on tasks humans excel at.

Excerpts & Key Quotes

Limited cognitive capacity

"Our intelligence is therefore relatively high compared to other animals, but in absolute terms it may be very limited in its physical computing capacity, albeit only by the limited size of our brain and its maximal possible number of neurons and glia cells, (e.g. Kahle, 1979)."

Comment:

We are often unaware that AI is still in its infancy. Considering that we humans may be limited in our physical computing capacity, it might be the case that AI will surpass us eventually. When we see intelligence from an anthropocentric perspective, our own limitations become unimportant in determining intelligence. Our limitations are not characteristics that make us humans intelligent and thus not part of our general definition of intelligence. In doing so we fail to see the potential for intelligence in AI. Limiting AI to this anthropocentric view of intelligence instead of enabling its own. This does not necessarily mean we are not as smart as we think. We could still be very intelligent in our strengths. After we have separated different kinds of intelligence human intelligence could still prove to be impressive. The human brain architecture could enable human intelligence to be a strong competitor for AI architecture and intelligence. However, the antropocentric view on intelligence is shown to be outdated.

Ingrained cognitive biases

"Our limited processing capacity for cognitive tasks is not the only factor determining our cognitive intelligence. Except for an overall limited processing capacity, human cognitive information processing shows systematic distortions. These are manifested in many cognitive biases (Tversky and Kahneman, 1973, Tversky and Kahneman, 1974)."

Comment:

Where the text earlier argued that our limited processing capacity would make us less intelligent than we often assume. An argumentation I agreed with up to a certain point. It now argues that our cognitive information processing is not only limited but also flawed because of ingrained biases. We often make 'irrational', 'intuitive' choices that are based on biases that are often unconscious and implicit. This can result in suboptimal decision making, given the decision time and available information. Consequently, we really do seem less smart than we think. While I agree that this conclusion, I would also argue that is seems to be a bit unfair to humans. Yes, we have ingrained biases and these play a role in unconscious decision making. However, we are also aware of them very often and try not to base every decision on our 'intuitions' alone. Which to me shows we are able to combat our biases to a certain extent. That to me also shows a certain intelligence.

Multiple Narrow AI is most relevant right now

"This includes, for example, a proper understanding of the basic characteristics, possibilities, and limitations of the AI’s cognitive system properties without anthropocentric and/or anthropomorphic misconceptions."

Comment:

While the anthropocentric misconception get a lot of attention in this text, anthropomorphism is only mentioned ones. However, it seems important for their framework on Intelligence Awareness training. As solely contributing a more appropriate level of intelligence does not enable us to utilize AI appropriately. Understanding how it think and how this differs from our own way of thinking seems crucial. Exactly why it would have been interesting to see more explanation and examples of anthropomorphic misconceptions.