innovator

Add idea


Calendar

«    February 2019    »
MonTueWedThuFriSatSun
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
 

 

Advert

 

Payment

 

Advert

 

Authorization

Стартап
DateDate: 12-02-2019, 06:24

Artificial intelligence can play chess, drive cars and look for new medicines. This type of AI is called limited or weak. These are just advanced algorithms that perform specific tasks. Their abilities and scope of application is constantly growing, but their danger to humans is limited by the loss of work, reports HighTech.
The new generation of AI will certainly raise the stakes, scientists say. The combination of computing power of machines and human intelligence will allow a strong AI to learn, solve complex tasks and improve themselves. And to engage in tasks for which no one prepared them. So AI can turn into an artificial super-intelligence (ICI). According to some experts, its appearance is possible in the period from 2029 to the end of the century, according to Science Alert.
Whenever this happens, there is no doubt that the ISI will change humanity. Perhaps he will solve the world's problems of lack of resources or climate change.
But the lack of proper control will lead to disastrous results. And it's not about killer robots.
IISI is unlikely to be hostile to people. But the car can come to the conclusion that the most effective strategy to solve the problem that people have assigned to it is to get rid of some people.
Here are the possible scenarios. ISI, which develops a cure for the virus, decides to kill all its carriers. The military drone comes to the conclusion that the only way to ensure the security of the allied forces is to destroy the entire settlement along with the civilians. A neural network, designed to protect humanity from climate catastrophe, erases from the face of the Earth all the technologies that pollute it.
Now, such layouts seem fantastic, but even on a more modest scale, they can cause a lot of harm. To avoid such scenarios, scientists propose to introduce three methods of control.
Source: NEWS.am





Add comment

Your name:*
Your E-Mail:*
Input code
from the picture*
Enable this picture to display the security code
reload, if the code cannot be seen