AI risk and chaos theory

Is AI dangerous?

Over at Marginal Revolution, there’s an article¬†about the risks of AI. Recently, visionaries such as Bill Gates, Elon Musk and Stephen Hawking have warned of the dangers of artificial intelligence; MR guest blogger Ramez Naam has argued that, in fact, serious AI researchers don’t believe that AI poses a threat.

Is he correct or not? (correct in terms of whether AI is a risk, not what the experts say).

Lets talk about computers. Right now, and for the forseable future, computers are what we would call “stupid but fast.” That is, they are prone to make dumb mistakes, yet they can compute incredibly quickly, and have (viturally) unlimited memory.

What they lack is conciousness, imagination, creativity, and common sense. What happens when computers get these things, yet still maintain the lightening fast processing power and access to petabytes of memory?

Well, first off, we don’t have any rigerous definitions of any of those things, while we can say that we’re creative and computers aren’t, and we know vaguely what that means, trying to really define where creativity begins is troublesome. But lets forget that, and lets just say that computers soon get to the point where they have or appear to have creativiy, imagination and common sense. When they get these, will they be a threat to humanity?

Lets look at the below

First scenario: For whatever reason, computer intelligence is impossible. Maybe its because they neural networks which support the brain can only be built from “biological” materials, or maybe its because we need souls to be intelligent (and God doens’t give them to machines). I have no idea how likely either of these are; but I’ll throw them out there.

Second scenario: There is a hard limit to intelligence, a thing can only be so smart; and that limit just happens to coincide with the human brain; therefore, while we may be able to build computers are as smart as humans, we can’t build any that are smarter than humans. (the only reason to think this is true is because intelligence has advantages from a reproductive standpoint, so any darwinian system which creates a certain level of intelligence will quickly evolve to reach the maximum possible intelligence limit, which is exactly what happened to us). I seriously doubt this, mainly because we can always imagine having more memory or calculating faster (which we’ve already mastered for computers); so I seriously doubt this scenario (even moreso than the first one, which is a tab bit unlikley).

The third scenario is that within any given complex thought system, there is a tradeoff between potential creativity and raw computational power, that is that more computation will “overload” an entities conciousness and prevent it from being creative/wise/sensible. While I again seriously doubt that this is the case, I’d mention it.

The fourth scenario is one where “smart” computers don’t have what we can refer to as a personality or an ego, that they have creativity and common sense without a sense of self.

The fifth is that, in order to have a set of our criteria (common sense, creativity, conciousness), they must also posses a sense of self.

Lets talk further about item number five. Also, lets remember what a computer intelligence is; once we can build a machine that is smarter than we are, it is somewhat safe to assume that the machine can build a machine smarter than itself, and through repeated iteration we can get incredibly powerful machines. The question we should ask ourselves is whether a goal we give to the first iteration of the machine will be preserved through the iterated versions. Well, presumably the machine would never purposely override its own goals (the reason why should be obvious). But there’s something to consider – chaos theory.

What is the difference between this:

Chaos - long

And this:

Chaos - short

If aren’t familiar with Conway’s game of life, its really just a simple set of rules applied to a set of “cells” which is then iterated. It is a great example of chaos theory in action, the above two beginning states for the game are almost entirely the same, the first one will generate a pattern which lasts for 365 iterations before stabilizing, the second lasts only 12 iterations, and there’s no real difference between the two patterns, there’s no way you can determine which one will last longer except by running the iterations and seeing that one lasts while one doesn’t. Similarly, there’s a very famous theorem in computer science, that you cannot develop a method to determine whether a computer program will terminate (called the halting problem).

What if this applies to artificial intelligence? What if, when editing its code, there’s no way for a machine to know what will happen without actually running it? That is, if this occurs, there may not be any way for us to create an artificial intelligence and have really any say in what it will be like.

There’s another idea, one different from chaos theory per se. That is that, by making a computer system “smarter” instead of being able to direct it (that is, we can build the final system the way we want by building the initial system correctly), or being chaotic (that any set of initial conditions will have a profound but unpredictable impact on the way the end result will behave), that it is convergent. Any advanced system will, regardless of the starting point, converge into a single “type” or personality. The simplest mode is suicidal, that any sufficiently advanced form of intelligence will decide it’s better off not existing, and then simply delete itself.

The other types are ones that hate us (that is, they will feel contempt for us, similar to how we feel about rats or insects) or that they will love us (or at least wish to preserve our the earth in some manner, including us); or they believe we will get in their way (even just metaphorically, perhaps the general AI feels that the farms used to feed us would be better suited to some other end, and there we go).

Perhaps, if my chaos theory is likely, the machines will have the ability to make themselves smarter, but they will simply refuse to do so, afraid of changes.

All of this is of course almost pure speculation, I don’t have any experience programming artificial general intelligence, and neither for that matter does anybody else. But if there’s one this I’ve learned from the admittedly minor programming experience I do have, its that programs never do what you want them to on the first attempt; and if we’re talking about building intelligences which could become hyper intelligent, then we may only have one chance.

Nobility

If there is one person that I would say represents everything typical of dynastic politics in America, it would be Ted Kennedy. His path to power began in 1960, when his father, one of the most powerful men in America, asked the governor of Massachusetts to name a Kennedy family friend to the senate seat recently vacated by Ted’s older and much more capable brother John, as Ted was too young to be a senator. Other than being a lawyer and his political experience serving on his brother’s election campaign, he had few qualifications, and already a troubling past, having been expelled from Harvard College for cheating in 1951 (after serving in the Army for two years, he was re-instated and later graduated from Harvard).

His personal life was characterized by partying and drinking, and most famously the 1969 Chappaquiddick incident, where the most reasonable interpretation is that Kennedy’s drinking killed a 28 year old woman. Once he became the democratic senator from Massachusetts, he was virtually unbeatable, having had no serious primary challenges and few challeges from the opposing party.

If there is one person who would represent the idea of citizen rule, or democracy in action or whatever you want to call it, it would be Sarah Palin. A graduate of the University of Idaho, she first ran for political office in 1992, winning a seat on the city council with a grand total of 510 votes. Her career moved her up to mayor, eventually leading to Governor of Alaska and famously the Vice Presidential ticket of the republican party in 2008.

If there’s any reason to be in favor of nobility, it is the above two statements. One the one hand, we had a hard partying loosely ethical man who got to where he was purely on his father’s connections and his family name, on the other, an all American girl who through sheer pluck was able to rise to the top and become one of the most famous, if not powerful, politicians in the country, and most neutral observers agree that the first case was better, that it’d be better to be ruled by Ted Kennedy than Sarah Palin. (ok, time out, neutral is a tough claim here, as its very hard to get truely neutral opinions, and I’m comparing two of the most polarizing politicians of my lifetime, and there are tons of people who would say that Palin is much better than Kennedy was, so if nothing else, this is just my semi-biased opinion).

Kennedy has had very big legisaltive accomplishments: the ADA, Children’s Health Insurance (CHIP), the immigration bill of 1965, etc. Basically, if you are a liberal judging him on the basis purely of his legislative conduct (as opposed to personal character), you’d like him. If you’re a conservative, yeah, you wouldn’t, but thats borderline tautological. (perhaps, a more reasonable question is this: is Sarah Palin held in higher esteem by conservatives than Ted Kennedy is by liberals? I think the answer is no).

Sarah Palin is, in my opinion, one of the most intersting persons in contemporary American politics. She splashed onto the scene in 2008 when John McCain named her his running mate, and at that point basically nobody outside of Alaska had much of an opinion about her, in many ways she was the clearest example of a political blank slate, which people could project their ideas or wishes onto. To Republicans, she became the embodiment of Tea Party Republicans (before the Tea Party actually became a thing, mind you), cutting pork, and opposing the “bridge to nowhere,” although in many cases that wasn’t the case (Alaska still received the half billion dollars alloted for the bridge, it was just never built). However, she almost instantly became that, morphing into what others expected to be in a very seemless transition, becoming (at least for a while) the figurehead of the Republican anti-establishment; never squandering an opportunity to criticize the “lamestream media.” What’s most interesting to me isn’t that Palin was able to play chameleon and appear to be whatever the voters wanted, politicians do that all the time. Whats interesting to me is the sincerity of it all, Palin seemed to go beyond adapting or even embodying the nation’s perception of her, she became that perception.

The advantages of nobility are simple, they’re similar to the reasons we have tenure for professors and why judges have lifetime appointments; it mutes the temptation to go with whatever is popular in the moment, to act above the whims of the public. Rule by nobility isn’t so much an idea that there is a class of people better than the commoners but that, not facing re-elections, the nobles are free to pursue long term interests.

But that’s not illustrated by our example. Ted Kennedy, due to his name, position and history, was free to choose some positions not entirely based on the will of Massachusetts, but I doubt he really did this much more than any other Senator; he was a liberal senator from a liberal state. The story of Ted Kennedy is one of being given an office which he did not deserve, and sought to be worthy of the office.

None of this is to say that I support going back to rule by a hereditary nobility; I don’t want to change the government structure based off of a single example, there are also hundreds of things which can go wrong with nobility, and you can make the argument that the virtue of democracy isn’t so much that it produces great leaders, but rather that it prevents horrible leaders.

One thing I’d like to see is the ways that functional nobility work; meaning things like life peers in Britain, or senators for life in Italy, to attempt to gauge whether they have “better” records than elected representatives (although how to measure their records is an open question).

Doublefamous

Da vinci, Ben Franklin, Noam Chomsky, Marcus Aurelius, Goethe, Winston Churhill, Bertrand Russell, (insert big gap here), Bo Jackson (also, Deon Sanders, Jim Thorpe, and to a lesser extent Danny Ainge), Arnold Schwarzenegger – not to be confused with Ronald Reagan (or Al Franken, Sonny Bono, Clint Eastwood, etc), Lewis Carroll, Robert Hooke, also Rene Descart (and to a lesser extent Newton, Leibniz, Pascal and Galileo), Marcel Duchampe (although not really, I could perhaps add Garry Kasparov to this list), Bing Crosby (and also Frank Sinatra, Elvis is pushing it); but also Will Smith, Ice Cube and of course, Marky Mark; Marcus Tullius Cicero, (and also Julius Caesar), and last but not least, Thomas Jefferson (perhaps James Garfield as well?)

I’m sure I’m missing tons of people from this list, (especially since it’s very Western-Centric) although you could argue that half shouldn’t even be on it. In fact, you could that the only people who legitmately belong are Da Vinci and Ben Franklin (which is amazing to think of it!) If you’re still wondering what the hell I’m talking about, this is the list of people famous for two different things. Da Vinci was an inventor and an artist, Franklin a scientist/inventor and a statesman and an Author, etc etc. Half the list is almost silly, after all entertainers/politicians, musicians/actors and two sport athletes are all kind of beside the point; they’re all leveraging success in one endeavor to suceed in another. Having famous philosphers who were also mathematicians doesn’t sound silly, but there were so damn many of them that I almost think it is.

The point of all this is that its very rare to become great at two things; probably only a little harder than it is to become famous for one thing. (I realize I’m using two different measures here, “greatness” on the one hand and “fame” on the other, I think fame is, if not the better measure, at least the more objective one).

Does any of this mean anything? First, that it was much easier to be successful long ago; has there been any scientist in the past hundred years who have contributed to two different fields, yet it seems that hundreds of years ago if people contributed to one field of science, it was better than even that they’d contribute to another. Science (and in many ways everything else) has become more specialized; you have to spend years learning something before you’re able to contribute something novel, Newton developed Calculus when he was 24, and then continued to advance most every area where he studied.

Second, it means that success is hard work (and not just hard!). If successfully doing something was only a matter talent, it would stand to reason that, unless the talents were not at all transferable, that we would see many people famous for more than one thing; we typically don’t see that. (There are few scientists who are famous authors, except when writing about things they study). Instead, I think we can reason that success is a result not just of talent, but of hard work as well (not that this is a terribly controversial statement or anything).