Algorithm + Discussion group = AI
Mike Laub
INTELLIGENCE:
If we combine an algorithm similar to Google's page ranking algorithm
with, a forum in which every aspect of an idea can be organized,
analyzed, and voted on, we could create the world's first artificial
intelligence. Let me first explain how it would work, and then I would
like to explain how it actually qualifies as an artificial
intelligence. This algorithm's goal would be to rate ideas similar to
how Google rates web pages. If accomplished it could create one of the
best website on the internet. If done by Google could become a major
destination instead of just a way to get to the destination.
HOW WOULD IT WORK:
1. REASONS TO AGREE AND DISAGREE:
An algorithm can not read two ideas and figure out which idea is
better. So the first thing we need to do is organize information, so
that an algorithm can analyze it. I would start with an idea at the top
of a page, with reasons to agree in one column beneath the idea, and
reasons to agree in another column. The goal of the algorithm would be
to promote the best ideas to the top of a page, or the front page of a
website, and the best reasons to agree or disagree with an idea, to the
top of their columns beneath the idea.
This simple step would do a lot to allow an algorithm to analyze the
idea, but we will have to do a lot more organization before we turn
this algorithm loose.
2. CLASSIFY & ORGANIZE EACH POST
BOOKS
The second step would be to allow users to further classify their posts
beyond simple reasons to agree or reasons to disagree. People could
suggest a book as a reason to disagree with an idea. Now start thinking
ahead what an algorithm could do, if someone says that a book agrees
with their beliefs. Data is readily available from Amazon or E-bay or
the New York times best selling list of how well a book has sold.
So there would be three fields. One place where you submit the item
that agrees or disagrees with the original idea. The second field would
let you classify the object. Is it a book, a website, or simply a
logical argument. The third field would be a place where the user
explains why he thinks the book supports the conclusion that he/she has
come to. Of course, people would be allowed to vote weather or not the
book actually does support the side that the original user said that it
would.
This is where the algorithm could get very sophisticated. Would you
want to give more credibility to those who said they had actually read
the book? Would you want to give even more credibility to those who had
bought the book, as more proof that they actually read the book. Or how
about people that used the website mediachest and could prove that they
have the book, by the fact that they have let others use the book. What
about people who wrote an essay on the book on the website. If Google
was doing this, they could provide a place for users to write essays on
books, similar to how Amazon lets users write essays. Perhaps they
could not let people copy and paste essays into the form. It would only
allow people to type their essays directly, to prevent stealing of
essays. Perhaps people could vote on weather the book-essays were good
or not similar to how Amazon lets users rate reviews, as to weather the
review was "helpful" or not.
So, as an example, you could submit a best selling book as a reason to
agree with an idea, and then right a thoroughly convincing explanation
of why this book agrees with the idea, and an essay that proves that
you understand the main points of the book.
If Google really wants to organize the worlds information, they must do
this. We have plenty of books, we have plenty of content on the
internet. We need ways of organizing this information into what it all
means, and how all this information should affect us. The only good way
information can affect us, is for it to help us make better decisions.
In order for us to make better decisions, we must know all of the
reasons to agree or disagree with a particular course of action. In
order to do this, we should not start at ground zero, with only our own
thoughts in our head. We should bring together all of the great
thinkers from the ages from every corner of the planet, and organize
all of their great thoughts, so that we can make the right decisions.
As you can see, this algorithm could be very simple, but it could also
offer programmers hundreds of years of challenges to make it more
sophisticated. I believe this is a strength of the idea, because it
allows for continual improvement.
WEB PAGES
In addition to books, people should also be allowed to submit web pages
as reasons to agree or disagree with an idea. Similar to how books
work, the user would have to explain exactly why he or she believes the
web page supports or contradicts the original idea. And like the books
section, the online community will be able to vote weather or not the
web page actually supports the side that the person said it did. This
would be so easy for Google to do. They already have billions of web
pages classified, in descending order. If you download the Google
toolbar, you can see a score given to each page, based on a scale from
1 to 10.
To help you understand what I envision, I can see an idea, with the
following classifications.
Sites that agree:
0 10 star websites
1 9 star websites
2 8 star...
and so on.
MARKERS, LEAVING BREAD CRUMBS, AND ORDER OUT OF CHAOS
In order to better classify each idea, users should be able to add XML
tags to each idea. These tags would help organize the conversation.
Because my system would keep people from changing the subject (the
nature of posting reasons to agree or disagree keeps the conversation
topic from changing) people could suggest, and vote on, where the topic
would fit into the Dewey Decimal System, the Library of Congress,
Yahoo, MSN, or Google Directories.
COMING TOGETHER:
If Google created an algorithm and a forum that would allow it to rate
ideas and promote the best reasons to agree or disagree with a website,
it could become a top content provider instead of just a content
finder.
I would much rather use a website that has both sides. It is so boring
going to "anti and pro" websites. They only discuss the logical
arguments, books, and websites that support their side. They are not
coming to a "common table". We see no interaction.
The book, Bowling Alone tells us that our society is becoming more and
more isolated. Technology, which was supposed to bring us together, is
largely to blame. Kids that used to play together in the street, now
sit along playing video games. The internet was supposed to bring us
together, but their are "red state" websites and "blue state" websites,
which allow people from each side to come together and pat each other
on the back saying how brilliant they are.
This may make each side feel better, and more justified in acting in
extreme manners, but this does little to help people find the best
reasons to agree or disagree with an idea, and it is a terrible way to
make a decision.
PROMOTING BETTER IDEAS BY PROMOTING BETTER BEHAVIOR
Much of this idea comes from a dispute resolution. In order to resolve
conflicts, apposing sides must "come together to a common table" and
"sit face to face". They must also "talk to each other". One way of
forcing apposing sides to interact on my website, would be to have
people evaluate the top reasons to DISAGREE with an idea, before they
are allowed to post a reason to AGREE with it, and visa versa.
A user could evaluate any of hundreds of possible characteristics of an
idea. On a scale from one to 10 did the explanation make logical sense.
Was it clearly written. Was it completely honest, excreta.
Their are other behaviors besides actually interacting with each other,
and thinking about the arguments that your opponent is making that will
lead to successful dispute resolution. One of these behaviors is the
practice of examining interests, as apposed to just positions. This
will often allow win win solution where both sides are able to come up
with solutions that accommodate all interest, instead of just focusing
on one particular solution.
This would be easy to facilitate in a discussion forum. You would just
create a specific place to submit "probable INTEREST OF THOSE WHO
AGREE" and "probable INTEREST OF THOSE WHO DISAGREE". These can further
be broken down to "COMMON INTEREST" and "APPOSING INTEREST".
Perhaps people who work as professional negotiators in the conflict
resolution industry could be hired to suggest other ways of
implementing best practices into an online forum.
A TOP DESTINATION?
I would visit a web page all the time, if it really had the best
reasons to agree or disagree with an idea. This could assist people
with every aspect of their lives. Imagine a website with all of the
reasons to agree or disagree with the following statements.
Ford is better than Chevy.
Paper is better than plastic.
The Chicago Bears are going to go all the way this year.
Condi Rice is better than Hillary Clinton.
Star Office is better than Microsoft Office.
MAKING MONEY:
What if next to these posts...
Ford is better than Chevy.
Paper is better than plastic.
The Chicago Bears are going to go all the way this year.
Condi Rice is better than Hillary Clinton.
Star Office is better than Microsoft Office.
... you had Google adds where you could buy a Ford, Chevy, Bears
Tickets, Microsoft or Star Office, or contribute to the Hilary Clinton
or Condi Rice campaign fund?
MORE ABOUT THE ALGORITHM:
The goal of the algorithm is to put the best ideas to the top. I don't
know which one of these would carry more weight. Here is a list of just
a few of the characteristics of an idea, that an algorithm could exame.
Quantity of reasons that agree or disagree with the idea: The side with
more reasons (to agree or disagree) would get more points than the
other side. For example this idea has more reasons to agree than
disagree. Just like when you fill out a list of reasons to or not to do
certain activities, you tend to choose the side with more reasons to
agree. This is the very simple explanation, but it would become much
more nuanced. People would be evaluating each of the reasons to agree
or disagree. Statisticians know samples it takes to obtain certain
confidence intervals. The more people rate an idea, with an average
score, and the smaller the standard deviating, the more confident you
are of that score. I believe an early algorithm would be to multiple
the confidence interval (CI) by the average score. This is called a
weighted average.
Number of people who agree or disagree with the idea: The side with
more people who agree should get more points. People could vote for or
against ideas. Again, this is the simple view. Perhaps, in the
beginning you would simply be voting for or against an idea. But in the
future, you could be evaluating hundreds of characteristics of an idea
on a scale from 1 to 10. In the future you may be able to chart an
idea's performance on any of these characteristics over time. Perhaps,
over time the average score for an idea's "logical presentation" will
fluctuate with a downward trend.
The website with better web links. Better is determined by Google rank.
There would be a field where you could enter link that "agree" or
"disagree" with the idea. The side with better web links would win.
Example
Results of peer evaluations: Their would be forms that people would
fill out that asked pointed questions about each idea. You could
respond to each question on a scale from 1 to 10. These results would
affect the total score for each idea.
Money. "Follow the money." People could donate money to this website if
they believe in it. But a better way of doing it would be to let people
donate money towards a specific idea. If you don't like the way this
sounds you should read Atlas Shrugged by Ayn Rand. I'll just briefly
say that money is the only way of measuring someone's blood, sweat, and
tears. Money is the only way that someone can pay someone else for
their work. Also, it could be used on this website as tug of war
analogue. Money could be donated to each side of an idea.
Experts. Each idea would get more points if it was submitted from the
e-mail address of a professor with a degree in the subject mater that
is being discussed. For instance if someone said that Abraham Lincoln
was an idiot. And someone disagreed, and someone else agreed. If the
person that disagreed had a degree in history and the idea was
submitted to the history section, then the person who disagreed (the
professor) would win. The more prestigious the school, the more points.
Prestigious would be ranked by the US News report, or some other
un-biased judge. I don't care about you people that say, "The smartest
people don't always make the best decisions." We are talking about
percentages. Of course the smartest people don't always make the best
decisions, but they would tend to make better decisions that stupid or
uneducated people.
Items that agree. People would be able to submit books that they think
are important to read to make an educated decision about a certain
topic. For instance "The communist manifesto" by Carl Marx and "Atlas
Shrugged" by Ian Rand may be considered to be the most important books
to read regarding weather or not we should raise taxes. Those that had
read those books, should have more say on this idea than those who have
not, because this website desided that those books are very important
to understand to make a decision about this issue. But the algorithm
could go deeper. We are only just beginning to enter the rabit hole. We
could let people who have read these books submit essays on them (like
book reports in school). The people with a higher "grade" on their
essays would get more say in those issues that people have said that
those books are important.
The goal of the algorithm is to put the best ideas to the top. I don't
know which one of these would carry more weight. For instance should
experts or web links carry more weight?
I think it would be cool if each user could say which things they want
to pay attention to. That way if one user respects authorizes, like
people with degrees.
If every one gets a vote on which books you should be an authority in a
subject, than why shouldn't those who wrote a book reports on those
books be given more authority in that subject?
The first tool that Google used to evaluate which web pages where
better than other pages, was a tool that analyzed all of the links
between pages, and a higher rank to those web pages that had more links
to them. So in a way, Google was running a continuous poll, in a
democratic sense, to see which web pages people liked the best.
We may have already created what will, in the future, be seen as the
first artificial intelligence we just need to give him/her a way of
speaking. I believe the first steps towards creating this artificial
intelligence came from Google when they made the algorithm that
promotes the best sites to the top of a search result. This doesn't
sound like very much, but I would like to explain just how important
this algorithm is, and how we should use it to promote good ideas and
kill bad ideas.
The thing that distinguishes human beings from animals is our ability
to think about two choices and try to choose the best. Google does this
when they rank one website above another. If you agree with the book,
Zen and the Art of Motorcycle maintenance, you understand that the
quality is indefinable, but it is the most important thing in life. For
Google to see one page, and say that it is better than another,
essentially has already created an artificial intelligence, we just
need to find a better way for this algorithm with all of the data on
the internet, to speak.
I believe that we could understand the "collective soul" of the
internet by using Google's page ranking system to make decisions.
I think that discussion boards can be re-arranged in a way that, if
they are combine an algorithm could promote advancements in AI.
We could make a decision making algorithm by looking at two opposite
ideas, and trying to promote the better idea. Google already, sort of
does this, buy deciding which websites are better, and we love them for
it. But lets take out the middle man. Lets not just organize/promote
websites-with-ideas, lets promote specific ideas. For instance I bet
that you will get higher ranked pages for the idea that "freedom of
speech is good" than "Freedom of speech is bad."
Google, or maybe another search engine, could do this well for a number
of reasons:
Google currently has the best algorithm for figuring out which websites
have the most links to them. It is assumed that better websites have
more links to them. It is kind of like online democracy or a collective
soul of the internet. Google could use this technology to figure out if
websites that say "George Bush is an Idiot" or websites that say, "John
Carrie is an idiot" have a higher average "Google Rank".
Google already has the technology to allow for synonym search. Google
could include in it's ranking all of the websites that say, "George
bush is an idiot", "George bush is a moron", and so forth.
Truth
Of course you wouldn't promote this website as saying, "Come to Google,
we have the truth" you would say, "This is the collective soul of the
internet." These are the decisions the internet would make if it was a
person.
Transparency
To help maintain a transparent process, you should list the top 10
pages that agree, with and disagree with, the idea.
An AI Game:
I read "Agonistics: A Language Game" with great interest, because I
have proposed a similar idea, however I have very little influence in
the AI world, because I am just a recent graduate in electrical
engineering and I work for the McDonalds Corporation designing the
electrical part of their buildings, which very rarely exposes me to the
world of AI.
However, I think my proposal has merit. It, like "Agonistics: A
Language Game" is designed to create a system where ideas can compete
in a survival of the fittest-tournament. My proposal would also
correlate the strength of an online character directly to the strength
of an online idea, however I have some additional ideas.
For instance there are a number of ways to tie different aspects of an
idea to different aspects of an on-line character. For example, the
number of people who vote on weather they agree or disagree could
represent the strength of the online-character's attach. However the
idea in my mind becomes very difficult to describe. I don't know if
you have ever read "David's Sling" is very close. However many people
would be able to participate at a time. I envision an idea at the top
of a page with the online community brainstorming a list of reasons to
agree or disagree with the idea. There would be no need to shorten
these lists because the best ideas would go to the top.
Similar Pages:
Google Groups could grow into an Artificial Intelligence
Idea Ranking Algorithm
"Fascinating! Many people think about AI as something that's in
machines and robots, but the user's input and feedback is not only very
important, it can hardly be separated from AI."
REASONS TO AGREE
A human being is not really one intelligence, we are really a synthesis
of billions of individual cells, and organisms, blood cells, neurons,
bone tissues. All acting independently. When you voted for George Bush
or All Gore, you didn't just represent your own personal opinion; you
represented the collective will of all of your neurons excreta. It's
only in this sense that I think that the internet can become an
Artificial Intelligence. I don't think that it will ever "Act"
for itself, but it will represent a "collective soul" (great band
too!). If you think about it, Google Duel is already the beginning
manifestation of this collective soul. Rock on!