Comment on any post on autopilot and AI

Disclaimer: Giktaymes is a good platform to check how lonely you are in your mistakes.


Illustration from npr.org

Reflecting on the topic of IT popularity and influence in today's world, I came across very interesting information about the professions and training courses that are in demand today.
But first about thinking.

To believe that IT rules the world and it is IT that influences the development of the world - at least silly and not serious. IT is just a tool. People still, like many centuries ago, crave for wealth, and through it power, entertainment and feelings of envy for themselves. Thanks to IT, we just increased the mechanization of certain activities. In essence, an IT tool is all the same vocabulary and abacus, with the help of which we create devices and manage mechanisms.
')
People still control and influence decisions in our world. In order for IT to start affecting the world, artificial intelligence must appear. Moreover, such an intelligence that can live and survive, like any other creature lives on this planet. In my opinion, mankind simply does not need another AI, and is not clear by definition. In principle, there are no other examples around us. One can only fantasize that the whole Universe is some kind of rational and, possessing a set of rules and goals (perhaps even impermanent), somehow exists and generates new forms of life in itself.

AI will become real for me when I can create something that is now recognized as intellectual property. This is just something intangible that people can use. Whether it is the design of a car that has become popular among consumers, or the method of managing an enterprise whose products are in demand, and jobs are desired.

I want to draw attention to two skills: the first is to understand and use the “sense of beauty”, the second is to understand and use public morality. Those. The AI ​​should be able to write interesting articles in its blog, as well as express its opinion on the adoption of a particular state law. In the end, the AI ​​must decide to crush the pedestrian or allow the passenger to die. And do it yourself.

From my reflections, I see the main problem of both the development and perception of AI by man. The man behind AI does not see personal responsibility. If the AI ​​from some company makes a “bad” decision, then they will blame the company and this company is unlikely to be trusted further. But I can immediately argue to myself. There are thousands of surgeons in the world whose operations have failed for patients. But they still continue to work and the controversy over the “medical errors” do not subside. So maybe we can get used to the errors of AI.

In my reasoning, I do not consider the autopilot to be at least somehow involved in AI. Autopilot is the mechanization, the automation of a well-defined type of physical activity. The problem of introducing autopilot is the problem of limiting the spaces where the autopilot will work and where the person will operate. As in a factory where there is an automated workshop, workers are warned that being in this workshop can be life threatening.

Autopilot is just an automation, because its actions can be strictly formalized, and the results are specific and measurable.

And I think that the task of teaching the autopilot to ride among people is erroneous. The task must be to separate the autopilot from the people. For example, both trains without drivers, and long-distance communication without drivers.

Returning from what started. About interesting information. This information once again led me to the idea that those who proclaim useful goals for society always pursue something important for themselves. I keep thinking that business benefits are a side effect. And I ask myself: is everything really so simple in this business world? Just take the task, the solution of which promises a big profit. And then the opportunity is sought, hiding the true goals, to get the tools to solve their problems. The same autopilot is promoted as a life saving solution. And thinking about holy goals, thousands of noble programmers create the necessary software.

And looking at this infographic, I thought: maybe everything is really simple - it's just about the money.

There is simply an area where an employee for business is a problem asset. So why not automate this huge bag of money.

And what about possible other big pieces of cake?

Approximate total costs for employees (number of places * salary) can be estimated from statistics http://money.usnews.com , forbes.com , glassdoor.com and businessinsider.com .

As far as I studied statistics, the most money for people is spent in the following areas:

1. Medicine.
2. Programming.
3. Finance.
4. Engineering.
5. Jurisprudence.

But these are the areas where automation is most difficult.

So who's next on the automation queue? Most likely the one who stands at the entrance and exit of already automated systems. Automation will evolve. But the deep penetration of IT in the above five areas, it seems to me, is far from being charity on the part of the creators of the necessary IT tools. And among them, I would bet on finance and law. As practice shows in these areas more automated tools.

And by the way, from the same statistics it can be assumed that a person’s view of the world today is far from taking place through the IT sphere.

AI in these areas is unlikely to be directly involved. This AI will most likely be used in entertainment, politics and marketing. In those areas where you need to model the behavior of "reasonable" people. Those. I see psychology as the most probable field of activity of AI.

And the stories that the AI ​​will destroy people, because they are irrational - these are just philosophical thoughts of people about themselves and the problems of communication among themselves.

UPD. Wrote this article before I saw it today. Why isn’t the machine still doing this work? .