Διάβασέ το στα ελληνικά εδώ.
The father of Artificial Intelligence (AI), Alan Turing had developed the perfect test to determine when a machine can be considered “intelligent”: When the person interacting with it (written form of communication back then), cannot be certain whether he is interacting with another human or in fact a machine.
The last Microsoft publication, from Brand Smith and Harry Shum, titled The Future Computed, is dealing with the present and the future of Artificial Intelligence but not in a transcendental way as the usual publications. What that means is that it doesn’t delve into impressive future projections but rather examines the steps we are taking right now, the way the framework for the following steps should be shaped and what changes it will bring.
That’s because, for the AI systems to develop, it is necessary to safeguard the principles, the policies and the laws for their responsible use. In this publication the writers support the claim that these systems should be fair, trustworthy, transparent and controllable. They highlight though that before we make new rules and laws for AI, we should provide clear answers in fundamental issues concerning them. The same way that for the improvement of the systems, the authorities should ensure full data access while also staying on top of matters such as the safeguarding of individuality and privacy.
Changes will bring changes
The changes that the evolution of AI will bring are sweeping and will be affecting all of our life spectrum, starting from our places of work. Smith and Shum are researching in depth the future of labor not only as a field of study but also in terms of working positions.
Here are a few impressive numbers, which I will attempt to put in a brief order so that they make sense as a whole:
51,000,000 positions will be lost within the next decade but new areas of financial possibilities will emerge along with brand new fields and work positions. More specifically, until 2027, AI will displace 24.7 million work positions and will create 14.9 million new ones. That’s because every robot made for every 1000 employees reduces employment by 6.2 workers and causes wage reduction of 0.7%.
Yet technology has created many more jobs than the ones it has rendered obsolete, and it will keep doing the same in the long run. AI economy will present huge opportunities for employees and businesses alike, despite the initial adjustment period, so long as the societies respond to the new demands.
Easy? No. Possible? Yes. But our adjustment and response abilities will be put to the test. Mainly as to the extent to which we will be educated, not so much on AI systems but rather to adopt to the new circumstances that they create.
That’s where other existing data are also pointing to: By 2020 30% of the technological work positions will remain vacant because of the lack of talent, or, in other words, skill set. And this void will grow more since it will take time to develop the educational programs building those technological skill sets. Just imagine that 50% of the knowledge that a first-year university student accumulates will be completely useless until the graduation. Most employees will need to develop new skills during their working life.
If there is anything worth retaining from the above mentioned, concerning societies and individuals alike, maybe is the following:
The ability to learn new things, to cooperate, to communicate, to adapt to the changing environments, may be the most important skill towards constant employment.
Artificial General Intelligence
Since we examined the frame which must be developed and the demands which we should meet in order to develop AI systems in favor of our societies, let’s examine where we stand today in terms of design and operation of those systems, which is the main point of focus of Smith and Shum publication.
To better understand what it is exactly that we are talking about, we will use the definitions but also the differentiations that Max Tegmark provides in popular Life 3.0. As third level life he describes the one that has the ability to design its own hardware and software (technological stage). Contrary to us, humans, Life 2.0, who modify our hardware through evolution but design most part of our software (cultural stage). Life 1.0 is life that modifies its hardware and software only through evolution (biological stage), meaning primitive organisms.
That’s the stage we examine, the third stage defined as Artificial General Intelligence, meaning the ability of a system to successfully carry out any cognitive labor at least equally as good as a human would do. According to Tegmark, technosceptics believe we are still far from approaching that skill, unlike technology polemicists, digital utopians and the movement of beneficial AI who, despite their differences, think we are close to achieving that goal.
As we mentioned during the introduction, Microsoft’s research neither projects nor predicts. It examines where we stand at the moment and what spot we (plan to) give AI in our lives.
What we want AI to do for us
Three are the elements that played a part in the acceleration of AI progression the last few years: data increase and are more accessible, the cloud is getting bigger and the algorithms are getting stronger.
At Microsoft they treat AI as a sum of technologies that allows computers to perceive, to learn, to judge and help in decision making and problem solving, the same way humans function. The truth though is that today’s AI cannot even compete the ability of a child to understand and interact with people using its senses such as touch, vision and smell. Nor does it possess the basic skill to identify expression, tone, emotion and the subtle shades of human reaction. In other words, AI today has a high IQ but not much of an EQ.
When machines manage to incorporate the intelligence of IQ with the empathy of EQ, then we could say that we have managed to create a “communicational” AI. That is when it will be able to combine knowledge on different subjects and react naturally, perceive, for example, when someone is trying to be funny or sarcastic.
From InnerEye that is being used in Medical Tomography and uses AI technology, which was developed for video games, to Seeing AI that translates the visual environment for blind people, to Project Premonition where epidemiologists in order to prevent the spread of epidemics used mosquitos as sensors so as to collect and analyze blood, the kind of mechanical intelligence that computers can offer will have different impact in almost all fields where intelligence plays some part.
So, the aim is to develop systems which will be designed from the start to enhance natural human intelligence. That is what we need to realize when talking about the machines that will “replace” us.
AI seems to be the new solution to every problem. It is applied or called on for everything. Data emerge as the medium that will solve all of our problems. As a way to justify any means and every cause. As if we try to ignore our part in all of this. Thinking that this can help us get away with it….
We need you
As automation and AI take on tasks that need thought and judgement, it is of great importance for us humans to train ourselves, maybe by redefining some of our classical knowledge, theories and values, thinking critically, being creative and practicing empathy and logic.
Because, as Smith and Shum successfully point out:
Artificial Intelligence will be useful where Intelligence is!
Hearing Kevin Scot, CTO of Microsoft last August talking on Lex Friedman’s show about AI (you can find it on Spotify), it is easy to understand that Microsoft intends to do with AI exactly what it did with its famous products (software and hardware), to put it in every house, democratizing those systems. That why Scot states that they treat Artificial Intelligence as a developing platform, a platform that is open for the public to access, to manipulate and make its life better.
You see, Artificial Intelligence is not enough. It takes yours too…