Members of the Fake Intelligence scientific committee have recognized AI bubbles and stated them Fake Intelligence. These nine bubbles will be published one by one on this page, stay tuned.
Bubble no. 1: Everything can be solved with Deep Learning?
George Orwell introduces a concept of doublethink in his novel nineteen eighty-four. Doublethink means that person accepts two mutually contradictory beliefs as correct without any mental conflict or contradiction, i.e. without cognitive dissonance.
Deep Learning has brought a paradigm shift into AI programming. It has changed the way we see AI nowadays. Definitely one of the most important solution around AI.
Our doublethink around Deep Learning is that any intelligent operation can be done with it. No other thought allowed, or you are not a true member of AI society.
At the same time we know for sure (based on e.g. science) that human thinking is way more complex. We also know, that so far only relatively trivial cases are successfully implemented with Deep Learning. And still, as AI society, we doublethink that everything can be solved with Deep Learning.
PhD Harri Ketamo, Headai, Chairman Of The Board, Founder
Bubble no. 2: Every software is AI software
Artificial intelligence, data analytics and big data rose to public awareness some three years ago. As an immediate consequence, a huge number of software, IoT and robotics companies changed their marketing vocabulary. Now they are all AI experts in their business area.
Their web page once advertised their ERP software to support business growth. Now their artificial intelligence solutions optimally secure and guide the business growth. Likewise, they once used statistical methods to explore data, but nowadays they exploit data analytics to reveal hidden patterns. My very good friend Harri Ketamo (bubble 1), the major stakeholder of a real AI company, has portrayed the situation enjoyably: “If one has more rows in Excel than can be seen on the display, it is big data today”.
PhD Cimmo Nurmi, Satakunta University of Applied Sciences, Vice President
CI Computational Intelligence, Chairman Of The Board, Founder
Bubble no. 3: Optimization is now easy
For over 10 years ago me and my friend started to optimize. Only scientists knew how to. The ordinary man didn’t even recognize that it could be done.
Today everybody knows how to. Even those who doesn’t know they are in the business. They even seem to be the best ones in it. Almost.
PhD Jari Kyngäs, Satakunta University of Applied Sciences, Principal Lecturer
Bubble no. 4: Ai makes suspicious decisions!
We are afraid of artificial intelligence making non-perfect decisions. We are not worried who trains the AI and with what data. In fact we are not worried at all if we are fooled by beliefs told by other people, not even in the case when all the facts are against the belief.
AI is not figuring out anything on its own, it behaves based on algorithm and data. Both are people made decisions.
PhD Harri Ketamo, Headai, Chairman Of The Board, Founder
Bubble no. 5: Artificial intelligence changes how we learn in the future – or does it?
AI is changing quite a many areas in our society, but one thing is not changing – learning. And I mean learning in a very fundamental way. Learning is something that happens in ones’ brain. It is not yet fully understood how this actually happens, but it is known that our memory is associative. We remember things from associations. A smell from your childhood might bring up a memory of a place you used to live. Many couples have a special music that means a lot to them because it was playing when they first met. Playing the song again brings up nice memories of your partner. All this is happening inside our brain and AI is not changing that anytime soon.
However, AI will bring (and has already done so) new means to study and teach. It will even bring tools to monitor learning outcomes and even the learning process and can identify those students who are in risk of failing the course. In such, AI is just a tool for the learner and the teacher. While AI may also be used in providing learning content to the masses, teachers will still be needed to support and facilitate learning.
PhD Jari Multisilta, Satakunta University of Applied Sciences, President and rector
Bubble no. 6: Artificial Intelligence is Neural Network Intelligence
As the years go by, new buzzwords and industry jargon evolve and their meaning will change. Artificial Intelligence is a good epitome of this. The term AI truly emerged in the early ‘80s when the first academic AI conferences were held and the Lisp machines and parallel computing were introduced. In that time, AI was publicly advertised as rule-based expert systems. The expectations were high but the results were not that convincing. The public interest soon faded.
In ‘90s, old inventions were reinvented with the help of computer power increase. Neural networks (1943), evolutionary computation (1954), fuzzy systems (1965) and other nature-inspired algorithms evolved and started to show their abilities. A new term Computational Intelligence (CI) started to spread in order to cover all of these computational techniques. It was back then and still is debatable, whether CI is a subset of AI or vice versa.
Today, almost all AI publicity is centered on neural networks. This had led to the public conclusion that real artificial intelligence is obtained by using neural networks. Are the other intelligent algorithms then artificial artificial intelligence? The truth is that only a small number of current intelligent systems are based on neural networks. Significant number of celebrated intelligent applications actually use optimization algorithms designed in ‘80s and ‘90s. Still, the recent real-world applications using neural networks are very promising and convincing.
PhD Cimmo Nurmi, Satakunta University of Applied Sciences, Vice President
CI Computational Intelligence, Chairman Of The Board, Founder
Bubble no. 7: AI will generate jobs
Maybe one of the biggest bubbles we can create is about AI generating new jobs. Way too often people say that AI will generate more jobs than it will take. We have to understand that AI itself generates zero jobs. All new jobs are created by people and often enabled or powered by AI.
And vice versa, all decisions about giving peoples’ work to AI are made by people.
PhD Harri Ketamo, Headai, Chairman Of The Board, Founder