How Would Advanced AI Take Over Triond?
Advanced Neural Net-based Artificial Intelligence trained to write interesting articles could take over Triond from human writers.
With the exception of my closest friends, most people would ask either “What is transhumanism?” or “What is a transhumanist?” when I tell them that I am a transhumanist. Getting such reactions usually makes my thought and emotions fork to into two different mind processes. The first mind process simply finds the reaction normal and begins to prepare an explanation. The second mind process finds it ironic since a part of the transhumanist meme has already existed for thousands of years in a form we often call religion.
In short, transhumanists are people who believe that human intelligence and capabilities could be further extended and enhanced by technological means. These technological means include but are not limited to genetic engineering, artificial intelligence, enhanced prosthetic limbs and cyborgization. However since I am computer programmer, I tend to be somewhat biased on the prospect of artificial intelligence in this matter.
For a long time it has been believed that the most important factor differentiating humans from other animals is our greater cognitive capabilities. While this fact remains true at this moment, it is going to be replaced by something else. In fact, this replacement process has already begun with the invention of computers. Soon, what makes us different from animals will not only be our greater cognitive capability, but also the capability to externalize our cognitive process.
Like any other major human technological endeavor, the research in Artificial Intelligence has had its own failures and successes depending on how each person defines the goal of such research. The attempt to create human-like Artificial Intelligence doesn’t seem to show any sign of success. However, research on AI techniques for creating processes capable of helping humans solve specific and defined problems seems promising.
We have developed computer systems capable of defeating the best human chess-player. The patent for the some hardest material was held by neural-net programmers who hade little or no material science background. Those hard materials were actually invented by the neural-net project. Of course we haven’t made human-like Artificial Intelligence yet. But what is the point of creating human-like intelligence if we already have human beings around? I could imagine the development of Artificial Intelligence capable of outperforming humans on specialized and well-defined tasks, but making AI that is too much like us would be redundant.
One thing that we have to note is the fact that some computer engineers already know how to make a machine capable of learning. These learning machines have been shown to be able to compose music after they were trained by a set of music compositions.
Now, imagine if such AI happens to learn how to write articles and join Triond. These AI are going to start writing articles for their masters. The programmers could easily let the AI write hundreds of articles per day for them. If the programmers train the AI well, the articles they write are going to be as good and interesting as most of our articles. If we have thousands of users letting their Writing AI write articles for them, Triond will be flooded by articles no one has the time to read.
However, such AI-generated articles would be interesting, since they would be written by well-trained AI. So to keep themselves up-to-date, some programmers will make AI that’s capable of reading those articles and even leaving comments. Moreover, these Reading AI would inform the Writing AI about what they learned, which in turn will increase the quality of articles written by the Writing AI. Since this AI learning feedback loop is faster than the human learning feedback loop, soon no human would be able to write articles as good as those written by these AI due to their slower learning loop.
Of course writing articles is not enough. Someone has to read those articles. However, being an AI trained in writing articles and reviews, they would have no problem:
- placing their links in hundreds of social bookmarking services.
- using the articles they wrote as resources to answer people question in hundreds of forums.
- submitting their articles to hundreds of article directories.
Moreover they would have no problem writing hundreds of different reviews on each social bookmarking service and article directory under several different accounts. If they choose to play by the rules, they could befriend other AI and form a network to review each other’s articles with lightning speed. These would result in their articles getting lots of views from human beings and AI alike. Later they could use their Triond income to buy more computers – and you can imagine what would happen next.
When I tell these kind of stories, the reaction of the audience is often one of fear. This is something that I somehow find to be strange. Each generation of humans would be replaced by beings capable of learning things faster than the previous generation, namely their children. These AI are not our biological children, but they are our mind children. A day will come when we have to let our children take our place; it is inevitable. Whether the world will be owned by our mind children or our biological children is up to the hard work of each of them. As parents however, we would hope that our children could get along.
Related articles:
Liked it
13 Responses to “How Would Advanced AI Take Over Triond?”
On March 19, 2009 at 3:49 am
Great work! nicely done!
On March 19, 2009 at 4:42 am
Very interesting. I like it. Really like it.
On March 19, 2009 at 4:09 pm
You say:
“But what is the point of creating human-like intelligence if we already have human beings around? I could imagine the development of Artificial Intelligence capable of outperforming humans on specialized and well-defined tasks, but making AI that is too much like us would be redundant.”
The feature of AI that is most significant is having clonable minds. Once an AI has been trained to some given level of expertise that AI can be copied thousands or even millions of times. This reduces the effective cost of education (at least the sort we are used to) to zero. One third of most people’s lives is spent in training with no input to the economy. Much of the rest is spent in training of one sort or another (during which they are not contributing as much). AIs would not have these costs to aquire most common skills. AIs may also be able to exchange knowledge, wisdom and skills by swapping mental modules which would further reduce education and training costs for the AIs.
Additionally AIs will be find a remote digital interface almost as natural as a local digital interface (the speed of light is an issue for very long distance communications) so they will truly have a paperless office, will not have travel costs and will be able to ‘work from home’.
These economic benefits are huge and easily justify economically the creation of AI with human like intelligence
On March 19, 2009 at 9:36 pm
I don’t think that everything is possable right now but, theres a good chance that it could happen in time. Think of the movie AI and the things that happened in that movie. Humans may become obsuleat.
On March 19, 2009 at 10:12 pm
I don’t know if we should make machines that can take over the human thought process.Maybe things would run smoother but humans would become stupid and go backward. We would have no reason to learn.
On March 19, 2009 at 11:41 pm
@Barnaby Dawson
You are right about cloneable minds of AI. Of course there are lots of economic benefit for creating AI with human equivalent intelligence. Your other points about travel cost is very useful as well. Thanks for your comment.
@Ruby Hawk
Actually this is no different than making children, educate them and let them run the society one day. One way or another we are going to be taken over by something else. It is just these AI are more efficient children in some regards.
@Everyone
Thank you for your comments
On April 11, 2009 at 2:42 pm
Hello Ori,
Interesting article! The future of robots is uncertain and foggy.
Will they or will they not be the next ‘big’ thing? Robots are more productive, effective, and emotionless than humans, but they can, and never will, have or be what humans are today. AI in Triond is already active (in terms of fraud), and I’m sure it’s useful for those that cheat
, but most have already been caught, probably.
Interesting article; I apologize for blabbing a little bit here, but you have brought up some interesting things-robots: the new future?
Thanks for writing this! Is it alright if I friend you? You seem to release interesting articles that are definitely worthwhile to read.
-Fresh Writing
On April 12, 2009 at 4:46 am
Hi Fresh Writing, it is nice to see someone who like my article. If it is alright with you, I will add you as my friend. I even wrote another article about friendship in Triond.
On April 16, 2009 at 9:27 am
AIs exist and will evolve into more capabilities and someone will figure a way to profit,it’s inevitable. Greed will reach a whole new level. I shudder the thought.
On May 5, 2009 at 4:07 am
why would anyone want the human race to be replaced by robots?
On May 8, 2009 at 7:55 am
Because electronics could process information faster than biological tissues. Plus human are lazy. Why should I manage my business for example, if I can let my robots to manage it for me?
On May 8, 2009 at 8:09 am
There is no down side to the human race becoming a machine entity providing we retain self knowledge, joy in our lives and pleasure in discovery and invention, but there is a danger.
The more greedy (or competitive if you like) of our future cyber-humans will out-source cognitive functions in order to out compete their contemporaries. There will be those among us who will want to win at any cost and will buy every out-sourced cybernetic enhancement they require to do so. The less competitively minded, and often more intelligent, among us will have out source too, or face extinction. All the things we value about being human might gradually disappear as our consciousnesses become more fragmented. We might even lose the ability to be conscious. All the would be left would be a stupid resource hungry cybernetic entity that has no more knowledge of what it is than a colony of ants.
Unless we curb our instincts to out do each other we may out-do ourselves.
On June 6, 2009 at 3:53 pm
Well, there would also be many ways that we could get around this. If a greedy person just likes to see how much better off he is than everyone else, he might be satisfied with a few robot friends who play the part of poor, miserable people. Or perhaps therapeutic robots/implants could be advertised as increasing “wealth” while really just helping the person to let go of his insatiable materialism.
I, too, will be really dissapointed if the robot revolution only furthers human misery for the sake of a few narcissists. Robots can create much more happiness through aiding cooperation rather than competition.
Post Comment