The danger of artificial intelligence
The danger of artificial intelligence
The danger of artificial intelligence |
An experience I did where I wanted to from AI
Here's an experiment I did where I wanted AI to copy the paint colors to create new paint colors while giving the menu like the ones here on the left and here's what the AI actually came from. I thought I was ordering it like the names of paint colors. What I was actually asking is to imitate the types of character sets that I saw in the original text and I didn't tell anything about the meaning of words or maybe there are some words that should be avoided in these paint colors, so the whole world is the data that I provided
Like frozen yogurt seasons that no one else knows, so with information regularly asked by Amnesty International
This is a fish called Tensch and there was a group of researchers who trained in artificial intelligence to recognize these stress images, but when they asked her what part of the picture I was actually using to identify the fish here what is yes, the water hose is human fingers, so why are you looking for human fingers If she was trying to identify a good fish, it turned out that the ten chose the cup of the fish and so in many of the photos that Amnesty International saw about this shape, the fish looked like this during training and did not know that the fingers are not part of the fish, so you see the reason for the difficulty Artificial intelligence design that can in fact understand m It is seen and this is the reason why it is difficult to design images to identify self-driving cars and why a lot of malfunctions of self-driving cars are because of an artificial intelligence defect I want to talk about an example from 2016 there was a fatal accident when someone was using the autopilot automation system in Tesla but instead of using it on Highw I love the design it was because they used it on the city streets and what happened is that a truck got out in front of the car and the car failed to collapse now the AI was definitely trained to recognize the trucks and pictures but what seems to have happened I as a
manager when identifying the truck Come on For a highway where you expect to see the trucks behind the trucks on the side, this is not supposed to happen on the highway, and when you see the AI this truck looks like AI knows it is likely to be a road sign and therefore it is safe to drive under it, and here the AI spoiled from a different field Amazon recently had to drop the CV sorting algorithm they were working on when they discovered that the algorithm that had learned to discriminate against women was what happened. Did they train them on an example of the CVs of people who had hired in the past who are examples of that,
manager when identifying the truck Come on For a highway where you expect to see the trucks behind the trucks on the side, this is not supposed to happen on the highway, and when you see the AI this truck looks like AI knows it is likely to be a road sign and therefore it is safe to drive under it, and here the AI spoiled from a different field Amazon recently had to drop the CV sorting algorithm they were working on when they discovered that the algorithm that had learned to discriminate against women was what happened. Did they train them on an example of the CVs of people who had hired in the past who are examples of that,
AI has avoided CVs of a People who have gone to girls ’colleges or who have the word women like this somewhere in their CVs like in the women's soccer team or the Women's Engineers Syndicate do not know that Amnesty International is not supposed to copy this very thing that people saw and did what they asked to do Accidentally just to do the wrong thing and this is what can be intense all the time with ai ai ai can be really devastating and does not know him, so AIS who recommends new content on Facebook on YouTube is optimal to increase the number of clicks and views and unfortunately one way they found it to do this is to recommend the conspiracy theories or Intolerance, the AI itself does not really understand the content of this content and have no understanding of what the consequences of recommending this content are.
So when we deal with AI, it is up to us to avoid problems and avoid mistakes that may occur because of the old communication problem where we as human beings have to learn how to talk to them we need to understand what Amnesty International can do "to understand that Amnesty International, through its personality Small, Small, Really realizing what we're trying to ask to do it that way, YJ. A creative mind, We have to be prepared to work with human-made consciousness and that's the thing that we have so far and the man-made mental power nowadays strange enough Thank you
Post a Comment