Is Artificial Intelligence ready to Go?

Peter B. GiblettStarred Page By Peter B. Giblett, 24th Mar 2016 | Follow this author | RSS Feed | Short URL http://nut.bz/2szphs02/
Posted in Wikinut>News>Technology

Artificial Intelligence has now defeated world champions at many games, including Jeopardy, Backgammon, Chess and the Chinese strategy game Go. Does this mean AI systems are now fully developed and able to play a more meaningful role in the world? Should they?

Chinese Strategy Game

The ancient Chinese strategy game of "Go" has many more moves that must be considered for each each turn than does chess. Recently humankind was delivered another defeat by artificial intelligence (AI) when a computer called AlphaGo beat Lee Sedol, the Go world champion, so is it time for machines to rule the world?

Certainly not — while this now adds to a growing list of machines that have beaten the best humans at chess, checkers and backgammon. Truth is there is more to intelligent life that winning a few board games, interactions that manage society are very complex indeed, much more so than for a complex game like Go where it is true that two players having to look several moves ahead each having a gigantic number of possible outcomes yet this is only an example of dedicated AI at work, not general AI capabilities.

For each move in a game such as Go, the AI program will use a tree search in order to play out a series of scenarios, making a choice based on the most appropriate move and in truth computer programs are faster at sorting through a series of options and selecting the most effective path to take, it helps make computers good at winning games. Tree searches are a standard computational technique where there are a large number of possibilities available, it enables effective analysis of the choices. The first program I ever wrote was to play card games and when the computer knows all the rules it can be very tough for even the most skilled player to beat, with AI this is taken a step further as it analyses strategy and can indeed know the likelihood of the other player's next move.

The challenge with Go is that it is very hard to tell who is winning until the very end, making the tree searches somewhat challenging and most Go aficionados had considered it would be decades before a machine could come close to beating a human at the game, making this win all the more significant.

True AI is Something More

Creating an AI program that can win specific games is one thing, but true AI necessarily involves something more — things that can make a computer based intelligence useful in everyday situations. Yet at the same time legislators need to understanding that Artificial Intelligence could pose a significant threat to mankind in the future and need to if not managed correctly and need to be putting in place laws that force programmatic controls in all AI devices that govern how they work with mankind and how they treat people.

One positive example is the IBM-developed AI named Watson which in 2011 famously beat two former winners of "Jeopardy!" and since that time has aided lung cancer treatment programs. Additionally Watson's ability to process unstructured data was utilised by Woodside Petroleum as a way to integrate all of its information, from research reports to Twitter feeds. Each being methods that bring positive benefits, yet not all developers think this way and one aspect of an AI is the ability of a computer program to reprogram itself when it encounters a problem, on the surface this sounds good because the computer could test the most effective code and implement that, but this is precisely where societal dangers lie, where the computer uses logic to reach a different conclusion than human society may reach.

Humans may at times act illogically and irrationally at times, but they also have embedded in them a survival gene, which will help them do everything then can to stay alive and in the example of an airline pilot or bus driver keep others they are responsible alive. Machines have no concept of life or death.

AI or BI?

Of course it is argued that AI systems are already working with humans in many areas, including traffic management, airline scheduling, agricultural planning but it is questionable whether these are really Artificial Intelligence or simply examples of Business Intelligence at work as currently all final decisions are always made by human beings, with machines working within parameters set by knowledge workers who themselves have a great deal of specialist knowledge having been solving such problems for decades without the aid of computers.

Indeed this writer has been a part of providing Business Intelligence solutions for the majority of his career in information technology, and whilst such solutions are very intelligent, they are the result of good programming rather than a true AI at work.

The Need for Change!

Google and auto manufacturers have already called for changes to traffic laws to govern self driving cars and in reality this is only the start of many controls that are required.

But there are limits to how far we should be prepared to let AI rule our world and governments need to start taking action now in order to control these limits in the future, for example machines should never be allowed to set strategy. The three laws of Robotics, as defined by author Isaac Asimov as long ago as 1942 should be considered very much the starting point and machines must be able to distinguish themselves from humans and other intelligent life-forms that may exist in the future.

Image Credits

Images used here were obtained from the following locations:

  • Go strategy game by Agathoi CC0 Public domain from Pixabay
  • IBM Watson logo, courtesy IBM
  • Traffic Light Tree by Pierre Vivant in London's Canary Wharf courtesy Reuters
  • Three Laws of Robotics by Peter Giblett

Other work by Peter Giblett

Peter Giblett regularly publishes here on Wikinut and contributes a semi-regular column on Two Drops of Ink, a site dedicated to the improvement of writing, grammar, and prose and his own blog called GobbledeGoox. Recent works on Wikinut include:

Wikinut is great a place to share some of your own personal wisdom by adding a comment or becoming a writer, join Wikinut and write.

Tags

Ai Rule, Alphago, Artificial Intelligence, Business Intelligence, Chinese Strategy Game, Ibm Watson, Law, Programmatic Controls, Unstructured Data

Meet the author

author avatar Peter B. Giblett
Author of "Is your Business Ready? For the Social Media Revolution"

Social media consultant, with C-Level background.

Share this page

moderator johnnydod moderated this page.
If you have any complaints about this content, please let us know

Comments

author avatar Shamarie
25th Mar 2016 (#)

This is an awesome post, Peter! I was talking to a friend about this topic on the subway.

Reply to this comment

author avatar Peter B. Giblett
25th Mar 2016 (#)

As a former IT executive who is also qualified as a lawyer I find myself continuing to assess this question and know that our governments need to be protecting the place of humans in society.

Reply to this comment

author avatar Jennifer Dombrowski
27th Mar 2016 (#)

I loved this post. I had no idea of any of the info in this artcicle!

I always thougth you were brilliant but a lawyer, too? I wish i had half as many brains as you! But ia m trying to educate myself more! You give me the encourageemt to educate myself more! as i want to be a genius! Thanks for this article which i saw on Easter which made it that much better

Reply to this comment

author avatar Peter B. Giblett
28th Mar 2016 (#)

Any person can educate and improve themselves I wish you all the best with it.

Reply to this comment

author avatar spirited
28th Mar 2016 (#)

I enjoyed this article Peter.

I play both go, and chess.

It's a funny thing with chess and go computers, you can always beat them by allowing them to beat themselves.

When they tell you the best move to play against them which you then do, eventually I have found, you can usually beat them this way.

What I mean is you simply take back your own mistakes, and if you do this often enough you can beat them.

Cheating? No, just using their AI to help me beat them that way too.

Reply to this comment

author avatar spirited
28th Mar 2016 (#)

The point I was making above was that humans plus an AI, must always beat an AI, just on its own, mustn't it?

Reply to this comment

author avatar Peter B. Giblett
28th Mar 2016 (#)

Microsoft's Tay AI went terribly wrong this weekend.

Reply to this comment

author avatar Ian R Thorpe
11th Apr 2016 (#)

Very good article Peter, but I have to say the high frequency trading algorithms used by dealers in the world's stock markets over the past few years prove that even the most advanced business software has little to do with intelligence.

Reply to this comment

author avatar Peter B. Giblett
11th Apr 2016 (#)

Ian, Those algorithms have little intelligence about them, they assume stock prices will always trend upwards, but in a recession or market crash as we have seen on many occasions the price bucks the normal trend. There are mathematical formula to predict it, but most brokers are not interested in problems, they want quick profits and that acts to their peril at such times.

Reply to this comment

author avatar SaigonDeManila
3rd May 2016 (#)

I always believe it will come since Deep Blue was introduce o play compeetive chess with world chess championship n back in the days.

Reply to this comment

author avatar Peter B. Giblett
4th May 2016 (#)

Saigon, I believe there is a difference between being able to play strategy games really, really, well and being truly intelligent. The former simply demonstrates an ability to analyse and value moves the later is something more, much more.

Reply to this comment

author avatar SaigonDeManila
4th May 2016 (#)

Agrees but that was the symbolic faith behind the win of Deep Blue not just for a lousy cold zinc chess apps but to help other industries as well albeit without personality or emotion.

Reply to this comment

author avatar Peter B. Giblett
5th May 2016 (#)

Things have advanced a lot since Deep Blue. The catastrophe of Microsoft's Tay.AI is proving that machine - human communications are quite a challenge.

Reply to this comment

author avatar Abdul Batin Bey
29th May 2016 (#)

A great write and I think we humans should always tread lightly when it comes to A.I. but I also feel that we will see major changes in our lifetime. The inner Luddite in me cringes.

Reply to this comment

author avatar Peter B. Giblett
30th May 2016 (#)

I am the opposite of a Luddite, I embrace technology and have spent so many years developing it. But in my view we have not thought through all the possibilities with AI.

Reply to this comment

author avatar the fool
5th Jun 2016 (#)

I wonder if A.I. could ever have a curiosity factor built into it.

Most of our advancements are because we are curious for an answer.

We explore outer space to see if other life forms exist, for example, what interest would A.I. have in doing this?

Reply to this comment

author avatar Peter B. Giblett
6th Jun 2016 (#)

Sir, you are right that many human advances are because of that curiosity factor. AIs are still machines and have not yet developed feelings, I suspect that is a big step.

Reply to this comment

author avatar the fool
6th Jun 2016 (#)

Yes, if they developed feelings, they would perhaps be only interested in themselves then, or would they?

Would they still be interested in finding cures for the human's illnesses too, or in exploring outer space, as well as their improving their own functioning, and increasing their own "intelligence".

Is the curiosity factor just a feeling, or is it actually something more than just that?

Perhaps it is a trait of intelligence too, or maybe it's a "higher" level feeling.

Funny that you called me "Sir", given my pseudonym name here.

It shows you respect people, which to my mind is another great trait of humans, and it's maybe more than just a feeling too.

Reply to this comment

author avatar Peter B. Giblett
7th Jun 2016 (#)

Sir, truth is computers can be taught to respect people, it merely requires the writing of a subroutine to do it, albeit a complex subroutine. Truth is though not all people respect other people.

The search for higher level intelligence is an interesting question, I wonder if it is something we humans have worked very little on since the time of Plato and Socrates. That said us humans are not simply interested in human well being, plenty of us specialise in the well being of other animals (and perhaps more would do so if our society weren't so money driven).

Reply to this comment

author avatar vandana7
11th Feb 2017 (#)

Hello bro in law..waiting for you on myLot..

Reply to this comment

Add a comment
Username
Can't login?
Password