cool hit counter The Age of AI: Algorithms in the Sky, Ethics in the Ground_Intefrankly

The Age of AI: Algorithms in the Sky, Ethics in the Ground


With the rise of AI fever, algorithms, a term originally reserved for the computer industry, have also started to appear frequently in the public eye. On closer inspection, the words algorithm and arithmetic are rather mysterious. Algorithms were originally defined as methods and ideas used to solve problems in the computer profession. The term is a long way from the popular perception. But the moment AI started to intervene in each of our daily lives, we also started to be managed by algorithms and models.

Knuth, the algorithm god, author of the Art of Programming book series, Stanford professor, and Turing Machine winner, said that algorithms + data structures + programming languages = computer science. This pretty much illustrates the importance of algorithms to the field of computing. Of course, there is actually a difference between algorithms in the classic sense, and the algorithms and algorithm engineers that we all discuss today in the age of AI. But as a foodie, it doesn't matter what you know or don't know about this difference.

In the years we've been unaware of, algorithms have started to take over every aspect of our lives. For example, the articles you read every day are carefully recommended by the algorithm. There is a ghost of an algorithm behind all the information about your food, clothing, housing and transportation, just those things that are situated in the app of your smartphone. Suffice it to say that every aspect of an individual human's life is being taken over by algorithms.

Is this takeover good or bad? Some people fear that AI will destroy humanity, others feel that AI brings a better future to humanity. I think there may be no right or wrong technology, but the humans who use it are not so sure


There is a very famous group experiment in game theory. One version of this goes roughly like this. A group of people enter and each pay $100 entry fee to play a game. The rules of the game are that each person can pick any number from 0 to 100. People don't know what the other person has chosen. Once the moderator has collected all the numbers from everyone, an average will be taken based on the numbers everyone has chosen. The person who picks that number closest to 2/3 of the average out of all the people in the room will win all the entry fees. If there is more than one person, everyone splits the prize equally.

This game is a lot of fun. Because we all know that if everyone picks randomly then the average is around 50 and 2/3 of the average is around 33. So if one prides themselves on being smart, they would definitely go with 33. A smarter person would think again, and since everyone is likely to pick 33, the overall average must be lower than 50, so a smarter person would pick the lower number.

This experiment has been done many times before at major American universities. Two conclusions were drawn.

The higher the IQ of the subject, the lower the final mean

Experiments that are repeated multiple times will always have lower and lower averages.

This experiment is highly informative for studying many group-based activities, especially those with mutual feedback.


Because I am abroad, I am rather out of date with a lot of information about the country. About two months or so ago, on the recommendation of a friend, I downloaded the famous so-and-so so-and-so news end. This app claims to personalize and recommend the news that is best for me. I thought maybe this would be a good way to get a sense of what's going on in the country and also what kind of a person I am through this app.

When I initially opened the app, I was shocked at what I saw. Emotional man bites dog isn't even news now. Again, emotionally I am a very vulgar person. But after all, I'm a computer science student. I know that for any algorithm to be able to recommend effectively, it needs data. And I'm now logging in for the first time and it's a blank slate. The roundworm in my stomach couldn't recommend anything to me, let alone so-and-so.

So what so-and-so recommends is actually the average within the entire current population, or within a certain segment of the population based on my personal information. That would be interesting. Turns out everyone is actually pretty cheesy. Heh heh.


We talked about that famous experiment in game theory. So can we assume that so-and-so's recommendation system, too, is a group activity. Can we further assume that this group activity converges in some direction if its objective function converges after how many repetitions.

That game theory game, if we change the winners to 1.5 times the average then the average will end up higher and higher. If we say the average wins, this game theoretically has the same result as repeating it as many times as once. In this way, the directionality of the reward function given to those who participate in the game is very important.

Then if you look at the world as a whole, human material civilization has been greatly enriched over the years. The arrival of AI has made man even more of a bipedal mammal with clothes on his back. Don't you think it's a great enrichment that some years ago, to read the news, you had to go to the portal search engine and actively search there, now you just have to wait quietly to be fed by so and so and so?

But we also cannot deny that the moral standards of humanity as a whole are declining. Our great country, for example, may not have been quite conceivable to people 20 years ago as it is today 20 years later. What exactly is causing the decline in morality is a very complex issue. But in terms of the evaluation function, what kind of things can be rewarded for doing, then after such group activities are repeated many times, the whole society is tilted in a certain direction.


So we're left with one last question, whether algorithm-led group activity brings us, an evaluation function of a soaring moral system or an evaluation function of a moral landing. Let's not get into the reasons for this one. At least it's obvious from the results. Why else would I see such a sight when I open so-and-so?

As for the reason, I think the reason is actually quite simple. Because all the dominant AI players who have mastered algorithms and computing power are companies that want and desire to make big bucks. We can't expect them to be thunderbolts. So the main point of their incentives is definitely that they want people who use their services to stay in their circle as much as possible. Ideally it should be 24 hours a day 365 days a year without leaving.

And the incentive that allows humans to survive is most certainly not high morality, but more the satisfaction of human needs and desires. As for the latter, it's both easy to make money and proven to work. What's not to like? So, we can see so-and-so and the like, the news is getting hunted more and more man-bites-dog, while its revenue and market capitalization keeps soaring.

My conclusion is that in the age of AI, algorithms go to heaven and ethics go to earth. Let's hope I'm wrong.

1、Blockchain has become the new windfall of the postInternet era is irrevocable
2、Another area where selfdriving technology is gaining popularity war
3、Photoshop is almost 28 this year I didnt realize you were such a Buddhist young man with the download of each version
4、New summer activities at the Shandong Science and Technology Museum are coming
5、Hangzhou mayor 2018 will see accelerated cultivation of blockchain second only to AI and virtual reality

    已推荐到看一看 和朋友分享想法
    最多200字,当前共 发送