When studying physics at the University of Zürich from 1999 – 2002 I also attended lectures about Artificial Intelligence (AI) and Philosophy.
Then, computer power was still far from mastering the necessary calculations. First of all, there had to be made decisions about WHAT exactly is Artificial Intelligence, even if it is around since 1954.
Years later, Robotics and Artificial Intelligence advanced at a crazy speed and even some of the highest barriers have been cracked lately.
Artificial Intelligence has become part of the weapons big players use to broaden their empire. Lots of investments are being made and systems are learning at a stunning speed.
Let’s Go Back To The Beginning …
A first step we learned/discussed in the lecture at the University was the following scenario:
A robot is acting inside of a rectangle space, moving from one side to the other. The obstacles placed in his way caused the robot to look for another way, without losing the original straight line of moving.
The great question was, how a robot could learn to take the same way back … far away from the situation where a robot should decide for his proper best way!
When watching the video below I was amused by the words of Demis Hassabis, about his first fascination about giving the computer tasks to do and it just solves them during the night. I share this fascination and remember well when my first program run for 14 hours and finished successfully
(It finally took 3 to 5 minutes to run, depending on the amount of data and it run 5 years in production until it was replaced)
I am far away from being a genius like Demis Hassabis and the result of my programming was by no means what Demis presented at a really young age.
Insights by a genius!
Demis Hassabis, founder of Deep Mind and now part of the Google team explains Artificial Intelligence and his genius steps in the video below.
Is Artificial Intelligence Bad?
Artificial Intelligence is not bad as such but quite dangerous in the hands of the wrong people.
When Albert Einstein understood Atomic Energy and made the knowledge available, the consequences have not been foreseen.
Is he responsible for atomic bombs thrown on Japan?
I hope the developers and project leaders at the different companies, driving the progress of Artificial Intelligence are aware of the relevance of their experiments.
I also have my doughts about the good ending as the acting of human beings is hardly driven by greed and fear.
If we teach computers Artificial Intelligence and they learn from us, why should they not be greedy and driven by fear?
Let’s Divide Artificial Intelligence Into 3 Steps Of Deepness
Artificial Narrow Intelligence
Definition: Machines acting based on the quantity of an endless growing database of collected and provided knowledge. This part of AI is mostly achieved through “Machine Learning”. Machines are competing with human intelligence on one specific field.
Opportunity: There are many possibilities for systems to help humans resolve tasks and leverage their skills. As Googles Deepmind concerning the GO game, we will discover solutions we have been blind about and it will solve many problems.
Danger: Automatisation through industry robots has caused the loss of many jobs in the working class. Automation together with Artificial Narrow Intelligence will cause the next wave of higher paid jobs loss. The good news is, there will also be many new jobs for everybody who is willing and able to reskill. Robots of any kind have to be trained and productivity might get to its peak.
Estimates are that 90% of all current workers will have to reskill and most probably various times!
But there are not too many ideas about the role of those who are not able to re-skill!!
The Six Figure Mentors provide a challenging solution to the problem.
Artificial General Intelligence
Definition: Machines are competing with human intelligence on most fields of intelligence.
The fact that the inventor of AlphaGo claims, that his solution can learn other fields of human intelligence without being reprogrammed means, that the next step is not too far away and AlphaGo is just a small project of Google Deepmind.
Even so other pioneers in the field are claiming that we are far from this expression of intelligence, but they have said so about many AI problems in the past.
The change of dimensions for this step seems logical to be far away but if we are looking at the progress of different schools of Artificial Intelligence and pair them up with current developments in robotics we might start to see the beginning of its development and it might be less far away than we think.
I imagine this would imply the cross discipline learning but also the replication of a humanoid body. Self-optimisation of AI
Artificial Super Intelligence
Definition: Machines are smarter than humans in any part of intelligence, including a self-defining Artificial Intelligence.
The definition implies some sort of Nirvana Status and it reminds me of the physical prediction of weather when chaos theory hit science in the 60’s. Before Physicists and Mathematicians believed that you could calculate anything through an eternal formula if you would just have enough exact data.
This great documentation illustrates the case and discovery of Chaos:
Chaos, its occurrence in all nonlinear processes and the acceptance of its existence changed all kinds of real world problems, but we also had to accept, the impossibility of calculating the universe.
What Is Creativity?
Is it the last asset of humans?
If a computer program like Google Deepmind chooses its steps in an unpredictable manner and changes our view on a complex game like GO, is this creativity or is a genius move just following a chaotic pattern?
Is creativity just part of the chaotic pattern?
Would a computer learn and discover ethics? Our ethics??
There are so many open questions … and scientists have just no answer!
I think there might be limits to the development of Artificial Super Intelligence like there is a physical limit to speed (light) and there is a physical limit to prognoses of weather, stock markets etc.
The Impact Of Artificial Narrow Intelligence
Scientists are working on the next step (Artificial General Intelligence) but Artificial Narrow Intelligence is already integrated into many of the devices and services we use daily!
We don’t need to make the step to Artificial General Intelligence to stay in front of a completely changed game! The ideas of Artificial General Intelligence or even Artificial Super Intelligence might be of the kind of ideas like landing on Mars, but even the small brother of them is strong enough to impact us heavily. All of us!
As the scientific protagonists declare, most of Artificial Narrow Intelligence is Machine Learning (ML).
Even this small step is frightening when thought through in detail.
Let me provide some examples and their implication:
This is real! By end of 2017, Elon Musk will send a Tesla of the newest generation from a parking lot in LA to a parking lot in NYC.
(and by the way, he is also developing a big truck …)
Yeah, that is what we are seeing as it concerns OUR roads!
But there have already been tests in different areas like trucks in mining or transportation of containers at ports or small transport units at Amazon Fulfillment Centers. They all can run automated by now!
Elon Musk looks forward to having self-driving cars on the roads in 2 years!
Google is the absolute “king of search” and they are implementing all kinds of AI into their algorithms.
If search algorithms know about the searcher, this has radical implication on what we actually see online and we have to be careful not to become blind for what is not presented to us.
It was a big surprise to see how Donald Trump got elected and nobody had foreseen it!
There is still a big part of citizens with no or very little online presence, not integrated into big data evaluations. But THEY ALSO VOTED!
I was quite shocked when Jay Kubassek presented input and outcome of a test of an AI service on a call on 11th of July 2017.
The analysis of 160 words led to a quite accurate profile of the writer!
In Digital Marketing we are trying to solve the gap between a target audience and the advertisement that corresponds to them. So imagine how this analysis will change the world of advertising. The big players like Google and Facebook already have access to these features for quite some time, but now some of these services are becoming usable to the mainstream.
Another field of exploration is voice recognition, where statistics exist about dictation being 300% faster than writing …
Analysis of Pictures and Video
The big changes of the last years in this area are stunning. They are responsible for one part of the Tesla success story.
Now, the vision of a computer is more or less 10 times better than our human eyes. The big challenges of recognition of shapes and their correct interpretation seem to be solved and in action with very good results.
Many AI programs are trained visually, so these techniques are of fundamental value and a step deeper into AI danger.
The big improvement currently seen is triggered by the ability of AI programs to learn through patterns (in
Some years ago a translator application could translate the words but did not understand phrases, so the outcome was quite unuseful. Through pattern recognition now entire phrases are analyzed and checked back with the AI brain to look for the best solution of interpretation and finally its translation.
This all is indeed a big help to solve many of our problems, even if it is far from proper machine intelligence.
The Big Players
There are currently different teams working on different solutions of AI, so in the future competition might erase some ideas and fusion others.
Currently, there are big players:
Besides these, there are many start-ups around the world.
Share – The New Age
Google by ethics is one of the pioneers of our sharing age with all its initiatives of Open Source support. The captured data is incredibly large and growing exponentially. This data, as well as the tools to implement AI, are available to anybody. But due to the mass of data and the complexity of calculations you need to have the capacity to hack into this field of science.
Everybody is talking about big data where we face similar problems. Individuals can hardly save enough of this data and be processing it, so it is bound to states, universities or big companies.
Like concerning Eduard Snowden and Julien Assange with his Wikileaks the philosophic discussion concerns, if the community of mankind keeps these
It is difficult to predict, where we are heading with AI, quantum computing, data mining, and robotics.
We can not avoid the development of Artificial Intelligence and it’s possible that it is the biggest impact mankind has ever seen.
Watching through many of these panels, no AI specialists, CEO’s and Professors can name solutions for the transition.
Please leave a comment if you know a good platform to re-skill for the digital age!
The solution I have found, who take leadership seriously in the transformation