01 September 2018

Trust in Science


Around us, many people do not believe in science. Some even go so far to claim science is merely some conspiracy or plot of government or big corporations.

Unfortunately, they partially have right.

Since the industrial revolution proved science works, various people and groups love to use term science to protect their interests, even in the cases when their interests go contrary to the science itself. Some of the most famous examples are fighting with the chemical industry to stop application of DDT, and the tobacco industry muddling of the scientific evidence. Today, the petrol industry is one of the main culprits, and also, you have loads of quacks and crooks that use fields of nutrition and medicine to grab a quick buck.
Such ‘scientific’ claims often have nothing to do with the actual science and the scientific method. The facts and conclusions used in arguments are usually full of logical fallacies and cherry-picking of the results. However, such organizations use term science and sometimes even pay people with Ph.D. to argue their point.
So it is not surprising that some people gave up on science; what they met as a ‘science’ is not science at all. And falsely, they concluded that all the science is like that.
Funnily, all this use of term science also proves the reliability of the science itself. Science and scientific method work, they are the best way for us to find the truth, and during the last thousand years both science and scientific method proved themselves. Otherwise, all those crooks would not use the term science to argue their point.
So, if one wishes to stand up for science, one has to explain not only the scientific method but also this muddling of the term ‘science.’ That is not as hard as it seems. The movement of skeptics calls this pretend ‘science’ a pseudoscience. And you know about pseudoscience. Not only that you heard about the pseudoscience, but most likely fell victim to some of the claims. So let us call pseudoscience with its actual name.
The reason why pseudoscience works is simple. People often have no time. Life is not easy for anyone. And so people take shortcuts.
It takes time and effort to check is something pseudoscience or science. The effort requires significant digging of the source material and reading loads of the stuff written in academese (that stiff, boring style, full of passive verbs and jargon words). And who has time to waste?
Luckily, there is a legitimate shortcut: if you’re not expert, follow the opinion of the majority of the scientific community. Science is genuinely a team effort, and the truth is uncovered only when a majority of the scientists in the particular area of expertise confirm that something is correct. One lone scientist claiming something opposite to the majority is suspicious. So if you are not expert in that area, ignore that lone guy. Remember the truth, reality, the universe, and everything are complex. There are no simple answers no matter how we wish for simple answers to exist. If someone offers you a simple answer, be suspicious.
And, follow your priorities. If something is significant to you, do the careful analysis, otherwise, go with the scientific consensus.


I will not hack on the example of the climate change, it is used over and over again, but instead, I will talk about something that is fun and obviously publicized because the term AI is now all hip and fashionable.
A few months ago of some dude decided to develop AI capable of teaching humans morality. And his solution is to point AI to humans and say do what they do. His project failed miserably because he never followed a scientific method.
Actually, seems like, he never bothered to actually look at humanity, read the books about humans and morality written over the past several thousand years, etc. Seems like he, also, never bothered even to pay attention to the errors in his own field of the Computer Science. He basically failed to complete the first step of any decent scientific research, reading and learning about previous, already established knowledge. 

The original concept is neat, AI that has morality, and we could use moral AI. But only someone who does not have expertise in sociology, philosophy, etc., would think that the right solution is just to point AI to humans and say do what they do. You see, Computer Scientists already tried that solution. They used humans to develop AI used in the justice system to predict if a criminal is likely to re-offend or not. And sadly that particular piece of software was used in real life. And guess what, it inherited all the prejudices and biases of the humans. So in the end, the software cleared murderers and flagged the persons with minor infractions. Because predictions used the color of the skin as one of the main factors. And I really mean murderers. A severe criminal was paroled based on this software and later committed another murder. But hey, he had a correct skin color.
The very concept of morality is culturally biased. Yes, we do have some basic morality as all other higher primates, but nuances of what is moral and acceptable vary not only from culture to culture but also within the members of the one culture as well. I did not research this particular issue in details, but the knowledge I collected so far was enough to realize that this dude will not be able to make moral AI that is not at the same time very, very biased and full of prejudices. In essence, his making of moral AI is simple pseudoscience because he never did a proper study of the previous research.

So, yeah, if you’re not expert in some field, stay with the scientific consensus, and any claim that goes against that consensus take with, not a grain, but a kilo of salt.


No comments:

Post a Comment