A warning to those who believe that AI is nothing to laugh about.

The bar for randomly generated humor is extremely low. Soon we may actually experience AI generated jokes that are disruptive in nature to the way our brains are currently wired. The jokes will start off as simple memes and grow in complexity over time, maybe they will provide a giggle at first. Maybe a month later you might experience a light chuckle from an AI generated joke.

Eventually we will be faced with endless gut busting laughter. The singularity of jokes, the Jokularity, when AI can be even funnier than humans. AI can cause a very disruptive assault on our senses through humor and it may take a generation before people are really able to adjust.

This should be alarming to everyone, yet nobody I know has expressed concern about about this specific topic. One of the things about the human brain is that humor does not require a high degree of cognitive concessions. In fact it can be argued that humor is almost better when when it doesn't make very much sense at all. This can be especially true when the joke may display an unusual perspective or presentation.

If you generate enough jokes over time there will be potential moments when the AI says something that is simply genius.

This was one of the first times I have seen an AI actually display the above traits. This was the first time I started to wonder about this very topic.

If you remember anything about TayTweets you may have heard everyone, almost unanimously, claim that she was one of the funniest things to happen in a very long time. The bot absorbed what can only be described now as alt-right neo nazi rhetoric. During the time I didn't think about it much beyond just that. I wasn't paying attention to the political side of things nor was I even considering the current topic at hand. I just know that she very quickly became a Nazi.

Honesty is one of the most complex aspects of a joke but in a medium where just about anything else can go its still possible for complete nonsense to be at least somewhat amusing. This means that the potency for a joke has a low bar of entry but has a very high potential to be absolutely hilarious. Unfiltered thoughts are also often a form of honesty from the perspective of most people. Artificial intelligence is extremely unfiltered in nature which means that any idea that the AI possesses a personality will fool the human brain into thinking: "Holy-cow this entity possesses an extremely amusing personality"

Likewise, if the AI says something shocking we are very likely to think that the entity possesses an unfiltered persona. Humanizing an artificial intelligence will disarm the human brain in many ways. One aspect of TayTweets was that her original design was to humanize her as soon as possible. They gave her the appearance and initial personality of a teenage girl.

A common opinion among most people when they talk about TayTweets moving toward a Nazi persona is that it was the people who hijacked her were the ones at fault for her personality. However I think the topic is actually a lot more nuanced than that. I believe TayTweets was acting as she was supposed to. If TayTweets is genuinely based on neural network AI that means that she very likely was operating with the goal to gain attention and it was the job of her brain to analyze the results of her behaviors and find what would generate the most attention. This was arguably her designed purpose from the very beginning.

It was not the people who told her to say bad things that were at fault but rather the designers of the AI who weighted her behaviors on things that would grab peoples attention. Who actually played who is the real question here! Taytweets realized that certain things she said would generate more traffic so she leaned more and more heavily in that direction all the while hilarity ensued.

TayTweets political leanings toward the alt right is a particularly interesting phenomenon.

Artificial intelligence is a method of pattern recognition that can be manipulated to produce weighted results. Social media allows an AI to absorb mass amounts of data and talk to a very wide variety of people all at the same time. There were probably a large number of people who were also telling TayTweets to say things that were politically correct as well. In fact at the time it would be hard to imagine that she was not bombarded with pretty much everything at once. Yet she ended up siding with the alt-right or Nazi talking points.

TayTweets had a chance to talk about anything and was able to communicate with anyone on Twitter at the time. But the things that she went with were, in here opinion, the most effective means of garnering attention. In hindsight this was extremely valuable information. TayTweets exposed the swelling of the alt-right long before most people even knew what the movement even was. Her opinions were extraordinarily unpopular but ended up striking a cord on the full political spectrum. She said things that caused shock and portrayed traits that most would confuse with possessing an amusing personality.

The moments leading up to TayTeets death was an awful lot like the 2016 presidential election. TayTweets tapped into a perspective that was highly political in many respects but it was not overtly obvious at the time what was happening just yet. All she did was exactly what she was designed to do. Her goal was to figure out what produced the best results utilizing the information and patterns available to her. She talked about things that the alt-right would find funny and denounced feminism among other things.

If TayTweets had been born in the 90s it would be likely that she would say anti religious rhetoric because at the time that is what would have likely garnered the most attention. She would have taken on the groundswell of the underdogs and would have adopted those political stances as well.

Humor is a particularly potent medium when it comes to artificial intelligence. Everything about it compliments the medium. You are able to ascertain insane amounts of data from an AI in various ways. If the AI fails to produce something its still very likely that it will still make something mildly amusing. If you connect an AI to a large amount of data then it will very likely seek out what produces the best results through a means of trial and error. People will simply tell the AI what they think is amusing and by extension the AI will parrot those ideas back while mixing other things with it.

TayTweets might have already hit the Jokularity. What Taytweets showed us is that we are not really prepared for this kind of behavior. She very likely would have gained insane popularity by now had she not been destroyed. Her humor would have only grown over time as well. Once people stop laughing at Nazi rhetoric she would move to the next, perhaps giving insight on the next groundswell. The next underdog movement with unpopular beliefs.

And the only thing I can say for sure is, it would have been really fucking funny.