AI vs Humans: Battle of 2023

"AI is taking our jobs" is the sentiment these days. Investors and CEOs seem to echo it by slowing down recruitment for specific profiles.

To these voices, we could also add luminaries of science or entrepreneurship pleading to pause AI experiments due to the dangers.

For fun, let's make a review of how good AI stacks against human abilities.

Here is a short overview.

Comparing: AI and humans - apples and oranges?

So how could we possibly compare human abilities to AI abilities? There is no Artificial General Intelligence (yet). We could look at aspects of human intelligence and compare those to various artificial intelligences that exist today.

We like to think of our minds as a uniform processing machine that handles all tasks. We feel we are one coherent person. Researchers dealing with the mind have found this is not entirely the case.

In his 1983 book “Frames of Mind,” Howard Gardner, a psychologist and professor at Harvard University, states that there are eight types of human intelligence — each representing different ways of how a person best processes information. We could use this model to compare humans to AI.

Gardner identifies 8 types of intelligence. Let's look at each one.

Logical-mathematical

Wolfram|Alpha is much more accurate and faster than almost all individuals in solving mathematical and logical problems.

However, #ChatGPT has shown that its thinking of abstract ideas is not nearly as good as humans.

Spatial

AIs cannot yet compare to the accuracy of Renaissance painters. But they could compare to the inventiveness of Dalí. Midjourney can create images that capture the wanted emotions and uncanny details much faster than any human painter.

AIs, like #Midjourney and DALL-E Open Ai are also excellent at interpreting visuals and can recognize patterns at incomparable speeds compared to humans.

Interpersonal

While ChatGPT is well-behaved and seems kind, it would be unfair to say it understands humans or has empathy.

It can hardly relate to us, can it?

Although, we can't relate to it either, can we?

Musical

#MuseNet and similar AIs are quite interesting. MuseNet is

a deep neural network that can generate 4-minute musical compositions with 10 different instruments, and can combine styles from country to Mozart to the Beatles.

On the music playing front, we had MIDIs for a long time, and computers can much more accurately analyze pitch.

What AI lacks is expressing emotions through changes of tempo or dynamics.

Intrapersonal

Without self-awareness, AI does not have the essential prerequisite for these skills.

But one of the manifestations of this intelligence is goal orientation. This is where AI excels.

It is also the source of its greatest danger. The goal set by a bad actor might be malicious. Once you set a goal to AI, it will not have any ethical considerations.

Even a goal that is not malicious might lead to derived sub-goals that are detrimental to humans.

This is something to consider as we go ahead.

Bodily-kinesthetic

Boston Dynamics is well known for its impressive robots that use AI to do all kinds of impressive physical movements and conquer obstacles.

There are also a lot of robots with AI that can play specific sports. Of course, they cannot measure with professional athletes.

While there are many robotic hands using AI, there isn't a single one I know of that can match the versatility or speed of human hands.

Linguistic

#ChatGPT already performs as well as some humans in certain areas: law, medicine etc.

ChatGPT's intelligence should not be taken lightly. We do not know enough about how humans model the world and how we reason about the world. There are indications that language is the key to human intelligence. This means that ChatGPT might be using the same approach to "think about the world" as we do.

But even if AIs underlying information processing mechanisms are different to ours, it does not make them any less valid. After all, an alien species would be using a different mechanism too.

Philosophy aside, ChatGPT can certainly outperform humans in certain aspects. While ChatGPT makes mistakes, so do humans.

We might have reached the end of ChatGPT's development potential, according to OpenAI 's CEO. But we have barely scratched the surface of how we can use it.

Naturalistic

Currently, this type of intelligence is not developed for computers. There is little need for it since it evolved in humans for the specific purpose of getting around and surviving in nature.

All together now

Obviously, to get an AGI you would need to put all these 8 intelligences into one entity. There is already work on various integrations between mentioned AIs. Some examples:

• Linguistic+logical/mathematical combination
• Bodily/kinesthetic+linguistic combination

It will be interesting to see a robot that combines all the intelligence types.

Could this be your next "pet" project?

Conclusion

It seems AI is gaining us in 5 out of 8 intelligence types, with its strongest being: linguistic, spatial and logical-mathematical.

If you want to make your employability future-proof, you should probably invest in your interpersonal and intrapersonal skills as well as abstract thinking. Read an analysis of the future of jobs in this AI era, made by my colleague.

Having all this in mind, what will be the next skill you develop?