Why you should fear Artificial Intelligence

“People fear what they don't understand and hate what they can't conquer.”  

- Andrew Smith, 1919

It’s a very common tendency that us, humans, share. And by that exactly we inevitably highlight one of our natural unconscious biases. With AI, just like any other complex matter, it’s rather easy and sometimes even fashionable to resonate with resistance more than in-depth research. Thus, I thought I’d started a series that can cover the main reasons why you should, or shouldn’t fear AI. 

Fear #1: AI will steal your job

To understand AI you need to figure out its basics

When hearing the term Artificial Intelligence, often referred to simply as AI, few associate it with algorithms and mathematics. 

The average person is tempted to automatically think of the universe of Matrix or Terminator, where an AI is a global threat able to conquer the world. This image in the shared popular consciousness forms our expectations regarding the evolution of humanity and creates scenarios and fears about the future that awaits us. 

The image shows the percentages of what ML is built of.

Observe the pattern

History has shown that progress has been constant and despite often being initially met with fear and skepticism, it could not be stopped or resisted, leaving the only possibility for it to be understood and embraced. 

Adapt or Die has always been one of nature’s most ruthless attributes.   

Revolution #1

Dating back to the first industrial revolution, humanity has managed to take a big step towards a better and hopefully more comfortable life among the masses, yet always with a little resistance.(see cartoon below)

With the creation of the first steam machine, jobs massively increased almost as fast as production levels. It reduced the influence of borders, making both goods and the transportation of goods easier and more affordable. 

Revolution #2

From the late 19th into the early 20th century, humanity has made a great leap forward again in technology, with the 2nd Industrial Revolution. 

With telegraphs and railroad networks, as well as now basic infrastructures such as gas lines and sewage systems, it opened up the possibility of more comfortable middle-class jobs and enabled cheaper and better goods that were transported everywhere faster.

Revolution #3 

With the development of the protocols used for internet communications in 1969, humanity was introduced to the 3rd Industrial Revolution. 

A new economic model arose, one of a sharing economy. The developing digital infrastructure created a more productive albeit more competitive economy, as competitors could now be located in even remote countries. 

This rapid and massive change did not go unchallenged, the spread of electricity has often been regarded with skepticism and fear. This can be seen in the following propaganda from 1900, where death seems to be the only result of involving electricity in every aspect of human life.

Revolution #4 (?)

Looking at the current speed of technological advances, we can assume that humanity is working towards the fourth Industrial Revolution. 

This time we talk about a digital one, which is a fusion between the mechanical and biological worlds.  

At the heart of it resides Artificial Intelligence, Neural-Technologies, Virtual & Augmented Realities. 

The change and its frequency on the technological side, has spawned a debate concerning the future of nowadays professions. The belief that Artificial Intelligence will cause a displacement of the human working force towards a robot-ruled one is one of the most frequently encountered on the internet, in books, in movies, or even in small cafes. 

Everyone is debating the future of the workforce, under the influence of the fourth Industrial Revolution, also referred to as Technological Singularity.

Will robots steal our job? | Is AI a menace for our workplace? | Are we going to be replaced?

Many other questions like this are rising, but in order to answer, we should take into account that AI falls into two general buckets: Artificial General Intelligence and Narrow AI.

Narrow AI

Narrow AI focuses on a specific, singular, or limited task. It is the type of AI that can be found in multiple applications in our daily life. For example diagnosis prediction or spam filtering, even shopping recommendations are some of many examples of narrow AI. 

The main characteristics of narrow AI are:

1. If trained with a certain set of data, it is expected to perform in a defined way, without being able to generalize.
2. They are characterized by a set of limitations and constraints.
3. They are led by a set of pre-learned parameters and rules. 

By introducing AI in our professional life, we can get rid of the most repetitive tasks which take up a big amount of our time. 

So, we are left with more time to be creative, to learn, and simply enjoy life.

If we take a closer look at the digital ecosystem, we can see that this type of AI has made a place in every sector of our life. 

Healthcare is considered one of the sectors which have gained the most. 

An example in this direction can be defined by the following AI-assisted surgery, which reduces surgeon variations that could improve the patient’s recovery. As it can be seen, the system does not replace the surgeon, but it aids him.

People in factories are assisted by robots, facial recognition systems are used for authentication and tagging, while companies like Netflix and Spotify use it in order to offer a better user experience, along with a faster sale. In my (humble) opinion, at this point, AI is developed horizontally and not vertically. In other words, there is an abundance of ML models which just enrich the Narrow AI spectrum. Adapting models is to various tasks is easier and more financially secure than developing a general one. 

Researchers tend to improve the already made systems which have the AI component. Despite being equal to digitalization, AI is a buzzword that is irrationally used in any conversation topic. 

Despite the best efforts from people from both the private and public sectors,

We haven’t yet reached true Artificial Intelligence.

Artificial General Intelligence

Artificial General Intelligence (AGI), also known as strong AI, represents a concept of a system with intelligence that emulates the framework of the human one, defined by emotions, beliefs, and needs. 

According to Sebastian Thrun, The CEO of Kitty Hawk Corporation:

‘Nobody phrases it this way, but I think that 

Artificial Intelligence is almost a humanities discipline. 

It’s really an attempt to understand human intelligence and human cognition.’

One characteristic of human intelligence is the ability to generalize, without being pre-trained, or in other words without prior knowledge. This is what researchers want to achieve with AGI. 

As previously mentioned, narrow AI must be trained with information, while AGI can leverage its knowledge by transferring it to new domains and tasks.

Another characteristic of AGI is the ability to self-learn and reason with its operating system, compared to narrow AI, where there is a need for fixed domain models provided by programmers. 

In other words, AGI doesn’t need to learn patterns, due to its ability to take spontaneous decisions.

Why does it take so long for AGI to come together? 

1. Lack of Funding

According to Dr. Ben Gortze, CEO and Founder of SingularityNET Foundation, the lack of funding for new trials of the AGI approaches is one of the biggest obstacles.

2. Business Balance

It is easier and more secure from the business side, to develop narrow AI, where predictions are made from already available datasets.

3. Resistance in Transition 

Gortze also highlights the fact that the narrow AI is characterized by such a multitude of models that the transition to a general artificial intelligence solution is rather challenging. 

According to various researchers, such as François Chollet, Google software engineer, or Jérôme Pesenti, VP of AI at Facebook:

The lack of progress towards general AI shouldn’t surprise us. 

Has anyone started?

DeepMind makes a fragile yet well-oriented step in this direction. It’s been outlined within their virtual playground through using reinforcement learning with entities (also called agents) that learn chess. 

The approach used is the one based on trial and error. By giving rewards to the agent, we keep him motivated to learn. 

Researchers from DeepMind are trying to develop a more General AI, which could solve a multitude of games, not just chess. But since we’re talking about chess…

Gary Kasparow, one of the world’s chess champions has managed to find probably the most fair and square approach about the future of jobs:

‘We have more to win than lose when it comes to AI. 

Rather than becoming obsolete, humans are going to be promoted.’

Kasparov also states that: 

‘Jobs don’t disappear, they evolve. 

Deleting people from repetitive jobs frees them up to be more creative. 

The future of the human race is hedged on creativity.’

Personal Takeaway

In other words, human intelligence is able to evolve along with artificial one. 

Going back to the questions above I believe that yes, some jobs may be taken by robots, but that’s a blessing, not a curse as long as there are professional transition programs. Yes, some jobs like truck drivers, factory employees, financial analysts, or construction laborers have decreased in demand. But AI along with technology is driving the creation of new jobs like Data scientists, Data consultants, E-commerce specialists, and so on.

Our species has the ability to adapt and evolve.

Observing the pattern, the percentage of new jobs for humans has all the resources and reason to surpass it, as exponential evolution is what differentiates this kind of ours as humans. 

So, choose: why should you fear AI?

1. for the unmet freedom it gives out for all the human creativity to reinvent itself and reality at once.

or

2. because you might end up loving it so much, you’ll regret being resistant to it in the beginning.

For as much as fear is an option, there are infinite more useful reactions towards AI. Evolution will happen, we just have to choose if we want to stand against it or if we embrace it.

Until next time,

Embrace evolution, factually.