Sunday, December 30, 2012

Internet Arguments.

In 1989, Mark Humphreys, a 29 year old student at UCD, wrote a 'chatbot' program called 'MGonz'.  He connected his chatbot to the university computer network and went home.

Overnight a student from Drake University in Iota, America was idly surfing the fledging internet and noticed the active connection from UCD and so being curious he requested more information about the user.

Only instead of the expected session information, the student received the following response from the chatbot MGonz: "cut this cryptic shit speak in full sentances".

The following day Mark Humphreys was stunned by logs of a 90 minute overnight heated argument between his chatbot and the student. The full conversation is available [here]

In 'The most human human' Brian Christain argues the success of MGonz was down it's abusive dialog which results in stateless arguments :

A close reading of this dialog ... turns up something interesting, and very telling: each remark after the first is only about the previous remark. The ... conversation has become stateless, unanchored from all context, a king of "marker chain" of riposte, meta-riposte, meta-meta-riposte.

All context was completely lost.  The issues under discussion were ignored.  Humphreys primitive chatbox worked because abuse is less complex.

These days the internet is covered with similar 'conversations'.  Comments on YouTube videos are nothing short of depressing.  Debates on political sites very quickly degrade into stateless and contextless point scoring.  Philosophical arguments very quickly lose all context when they meander into 'what does this term really mean?' territory.

All very depressing and all very human.  It's worth remembering a chief tenet of the Enlightenment was that the growth of knowledge was the key to human emancipation.  The birth of the internet from it's US military roots was heralded as rushing in the 'information age'.  Religion, dictatorships, economic slavery was all expected to die swift deaths in a world of free information.

But I'm wondering off topic.   Here I want to note the quality of discussion and debate on the internet is of the very low 'stateless' quality easily simulated by AI chatbots.  The Internet is a fantastic resource but its diet can be a poor one and it's consumption should be moderated.  The next time I'm tempted to post a rant, I'll try to remember I could just be arguing with an AI bot like *Ray Comfort.

*In order for me to say Ray is not an AI, I must have complete knowledge of the universe. I do not have complete knowledge of the universe so I cannot say Ray is not an AI (Readers of his blog will get the joke).


Post a Comment

Related Posts Plugin for WordPress, Blogger...