Within the recent decades, and even centuries undertakings done by humans have seen automation and machine replacements. The affected people and groups have subsequently transitioned to new occupations and the world has moved on. Such as when machines took over as the main driver within the textile industry during the Industrial Revolution (Wikipedia, 2016) or with 1700s breakthroughs in the field of storing images. Which eventually led to the first commercially available camera in 1826 (Wikipedia, 2016). When a machine is able to capture the moment one can argue that there is no more need of painters. Which to a certain degree is true; There’s no need to hire a painter to capture a family photo these days. On the other hand, the painters eventually adjusted to the new technology by accepting and embracing it and creating new fields such as abstract art, changing their career or continuing in the same path with photorealistic art alongside the camera.
The comparison to both the industrial revolution and how the camera became the AI of painters is translatable into the field of modern day AI. The term AI is the concept of machines being able to carry out tasks without explicitly being told how to do this (Marr, 2016). To break down AI further, AI is what is refered to as machine learning. Machine learning is about programming a computer without explicitly telling it what to do.
Suppose that you run an online store, you’re collecting customer information such as their age, gender, what products they visit, what the buy and other metrics. After a while, you have an ocean of records, by assessing machine learning you can use this data to try to predict what the next customer with a similar profile wants to buy and then present the next customer with ads or products suiting the customer’s taste (Marsland, 2015). This sort of machine learning is widely used in multiple applications where recommendations are made. Such as Netflix recommending a movie or TV show, or Amazon recommending you a certain product to buy.
Removing the magic of AI and Machine Learning
Machine Learning is a form of statistics, where one tries to make a model out of data, to later on, predict how new data can relate to earlier observations. Abstractly, let’s say we’re collecting data about someone driving a car to make this car drive by itself autonomously. The machine is collecting data about how the person is handling the steering wheel according to the road. When a 90° curve occurs the steering wheel is moved to adapt the wheels to the curve. This data is collected and the next time a similar curve occurs; the car will adjust the wheels according to previously seen data. Unfortunately, in a real word application, the process is somewhat more comprehensive and complicated, but the concept withstands.
Today, algorithms have to be fine-tuned and extensive work goes into each problem or task to be solved. There is a bunch of different algorithms, and no general way of applying one algorithm, on a broad selection of data. Each dataset has to be prepared in a different manner, and then an algorithm has to be fine-tuned in the best way possible make predictions on the dataset. This process is extensively human controlled, and there is no way of the AI to “escape” and suddenly self-develop in the same manner as evolution has formed humans and other creatures on our planet. Worth to mention is that AI can learn from itself and go into some sort of evolution stage with techniques such as Neural Nets and Deep Learning. These techniques try to mimic the human brain by letting the machine learn from previous tries. (Marr, 2016) On the contrary, this technique is still not making the AI truly smart, but it makes it able to process information based on earlier processes. For the AI to be truly smart we’ll need a new level of AI, a super intelligent AI that is self-aware and able to learn and develop itself based on experiences.
With recent headlines and famous scientists such as Stephen Hawking warning about the AI taking over (Cellan-Jones, 2014) it’s time to tone down this myth. Artificial Intelligence at its current step is quite primitive, we’ve seen great scientific progress within the field the recent decades and today AI and Machine Learning is a part of the daily life with personal assistants in our phones, recommendations systems online, Autonomous cars on the road and tonnes of other application out there. These applications are all advanced assessments of mathematics, statistics and computer science. On the other hand, they are still quite primitive. These are all tasks and calculations that a human with enough time can do.
When the AI becomes self-aware and adaptable
The wet dream of AI has to be the stage where it can closely mimic the human way of experiencing the world.
This type of AI is self-aware with the ability to understand it’s position in the world and from that point on underdo a state of evolution where it develops a world with itself as the centre. Additionally, it has all calculation power needed available, calculation power that vastly outperform our primitive human brain.
This form of AI quickly raises to the top of the food chain and the big question now is what will happen to its human fathers. This form of AI is nowhere in the horizon. The first step into this type of intelligence is to make a generalizable algorithm that can adapt to dissimilar problems without being explicitly programmed for each task. Additionally, a form of self-awareness has to be in a place where the AI can critically think for it selves without unconditionally doing as told by the human masters. With this form of AI still being a science fiction, the AI of today is taking over jobs and automating more and more tasks in our society. With the current progress, there will be no general need for drivers, delivery personnel or even shop clerks in the coming future.
Artificial Intelligence will automate jobs and in the same way it’s surrealistic to bring a painter around to capture the next photo of you and the lads on a night out; many of today’s whereabouts will have the same phantasmagorical feeling to it as more and more tasks are automated by our friend the AI.
Cellan-Jones, R., 2014. _BBC News. _[Online]
Available at: http://www.bbc.co.uk/news/technology-30290540
Marr, B., 2016. _Forbes. _[Online]
Available at: http://www.forbes.com/sites/bernardmarr/2016/12/06/what-is-the-difference-between-artificial-intelligence-and-machine-learning/#370963d1687c
[Accessed 23 1 2017].
Marsland, S., 2015. _MACHINE LEARNING An Algorithmic Perspective. _2nd Edition ed. s.l.:CRC Press.
Wikipedia, 2016. _Camera. _[Online]
Available at: https://en.wikipedia.org/wiki/Camera
[Accessed 23 1 2016].
Wikipedia, 2016. _Wikipedia. _[Online]
Available at: https://en.wikipedia.org/wiki/Industrial_Revolution
[Accessed 23 01 2016].