Tuesday, 6 September 2016

What we copyright, and how AI might change that

Copyright protects the expression of ideas, not ideas themselves. And at some point, artificial intelligence is going to automate the expression of an idea. This will mean that it will be possible to copy a copyright protected work without infringing copyright.

Computer games are a particularly good example. The graphics of the game are copyrightable as is the story. But not the gameplay. All the thousands of first person shooters out there don't pay royalties to the creators of Wolfenstein 3D. With advanced artificial intelligence it should be possible for the AI to analyse a game, understand the gameplay, and then create a new game that is sufficiently different in graphics and plot, but nonetheless captures idea. And given its the AI doing the implementation in the first place, what value then is there to the human coming up with the novel gameplay mechanic? Not much. There are cheaper clones within minutes of release.

To a lesser, but still significant, extent the same is true of movies and books. The plot can be analysed and an expression thereof created.

What does this tell us?

In some ways it suggests that copyright is somewhat strange: it values expression (implementation) more than ideas. And yet it is ideas that seem to me to be what advances us as a species. Someone could take the idea expressed in this blog post and, using different words express it themselves in a revenue-generating medium, and I would have no recourse against that person profiting from my idea (as an aside I have not done the necessary research to confirm this is an original idea, but I can confirm it is independent).

The other thing that this expected future suggests is that copyright will either be extended in some way to ideas, or will lessen in significance as automation of expression becomes feasible.

Optimizing train carriage utilization - signs on platforms

For various reasons different carriages in trains get fuller than others. This reduces the comfort of passengers, and in extreme cases means that other passengers do not have enough space to board.

But this is a fairly simple problem to solve with existing technology. We can measure the occupancy of a carriage in numerous ways:

  • optical cameras coupled with image recognition technology
  • thermal imaging that captures the number of separate heat sources, or perhaps the aggregate heat being emitted
  • thermometers would probably be sufficient to assess the occupancy to some degree
  • pressure sensors on the seats or floor
  • carbon dioxide levels
  • optical beams measuring entry and egress
  • etc
And we can communicate that information to passengers at subsequent stations. For example, the platform could have lights built into the floor which are greener where the carriages are emptier, and redder where they are fuller.

A couple of thoughts on The God Delusion by Richard Dawkins

P.14 “An atheist … is someone who believes there is nothing believes there is nothing beyond the natural, physical world, no supernatural creative intelligence lurking behind the observable universe, no soul that outlasts the body and no miracles – except in the sense of natural phenomena that we don’t yet understand”

By including the “except in the sense of natural phenomena” caveat Dawkins is effectively defining atheists as agnostics: he is saying there could be a god as long as god is a “natural” phenomenon.
It is interesting also that Dawkins’s definition is about belief. He is not starting from a philosophical position of what we know for certain and layering on personal experience, plus trust in the personal experiences of others.

It is also notable that Dawkins’s definition is a belief in the absence of something. In fact,there must be infinite things that Dawkins’s must actively believe don’t exist. He believes that Russell’s teapot doesn’t exist(personally I merely don’t know whether Russell’s teapot exists).

P. 31 “This book will advocate an alternative view: any creative intelligence, of sufficient complexity to design anything, comes into existence only as the end product of an extended process of gradual evolution.”

Dawkins’s does not seem to give adequate consideration to the possibility that a creative intelligence could have always existed. He has no evidence to rule out such a possibility. Given that our understanding of time as a dimension is extremely poor, I don’t think we can rule out the possibility of things that have always existed.

P.46 “The God Hypothesis is also very close to being ruled out by the laws of probability”

This is interesting in two ways:firstly, the God Hypothesis has not been completely ruled out, which implies agnosticism rather than atheism; and secondly, the book fails to provide any evidence on which to use probabilistic analysis to get“very close” to ruling out that hypothesis. He does provide his“Ultimate Boeing 747 gambit”, but that does not address the creation of intelligence by a mechanism we don’t currently know(other than random chance or evolution), it models to creation of an intelligence in our Universe with it’s natural laws, and it ignores the possibility of an intelligence that has always existed. All he does is use probability to suggest that the random formation of an intelligence is incredibly unlikely. He also doesn’t show it’s impossible. In fact we can use the anthropic principle to suggest that we are only here to speculate about the existence of a creator because, prior to the creation of the Universe because the forces before the creation of the Universe randomly created an intelligence that created the Universe (which itself is not improbable if time outside of the Universe is infinite – but as mentioned we(particularly I) don’t understand time).

P.46 Dawkins’s is somewhat scathing of agnostics, referring to the statement that they are “namby-pamby, mushy pap, weak-tea, weedy, pallid fence-sitters” as “partly right”. He says that “Temporary Agnosticism in Practice” is legitimate, implying that his other category of agnostics, “Permanent Agnosticism in Principle” are illegitimate.

In my opinion Dawkins is both wrong in is mistreatment of agnostics, and wrong in his division into TAP and PAP.

I would probably put my views into Dawkins’s TAP category but:
  • It may be possible that we may not be able to fully understand the universe – we don’t currently so what evidence is that that we will ever fully understand it? I’m not saying that we don’t, just that we don’t know that we will
  • Even if we fully understand the universe, might there be things that exist outside of the universe that we don’t understand? Perhaps our complete understanding of the universe will mathematically prove that nothing exists outside the universe. But perhaps not. If we are in a simulation, even if we completely understand all the physical laws of the simulation, will we get any understanding of the creator of the simulation
Hence my position is more like a Potentially Temporary Agnosticism in Practice, which begins to sound a fair bit like PAP. Although I’m definitely not “Permanently”anything as I realize new evidence could arise at any point.

However, this PTAP position is a day-to-day practical position where I believe in the reality I existing (which is expedient for day-to-day life). Philosophically my position is: that which I perceive exists, but I cannot know the nature of its existence. Philosophically I am firmly PAP, but not just about God, about everything(I know Dawkins exists as a concept, a name, but does he actually exist as a person? I don’t know).

The other aspect of Dawkins’s treatment of agnostics that I don’t understand is that given his extreme hostility towards religion, surely critically thinking agnostics are infinitely preferable to indoctrinated theists? Why does he go out of his way to alienate them?

P.50 Dawkins sets out a probability scale for God on 1 to 7, putting himself as a six. By putting himself at anything less than a seven he does, in my opinion, actually classify himself as an agnostic. But I realize this risk slipping into semantics.

But the other thing that’s interesting here is the information available to assess probability.As discussed above, I don’t think there is adequate evidence for probabilistic analysis. It’s not like assessing the probability of an earthquake in San Francisco within the next year, where we can look at records of previous earthquakes and other data on tectonic movements. There just isn’t evidence to justify where you put yourself on the probability scale.

If god exists outside of the universe then it is perhaps reasonable to assume that no evidence for or against its existence would necessarily be within the universe.

P.53 “The point of all these way-out examples [e.g. The Flying Spaghetti Monster] is that they are undisprovable, yet nobody thinks the hypothesis of their existence is on an even footing with the hypothesis of their non-existence”

Well Dawkins is definitely wrong on one count: there is at least one person (myself) that considers their existence / non-existence is on an even footing. Why? Because I have no strong evidence either way. Yes, you could argue that the fact that FSM was created as a parody indicates that it is not true. But then FSM could exist and have a very strange sense of humor. My point is that I don’t know, and that it doesn’t really matter either way.

P.56 “Why are scientists so cravenly respectful towards the ambitions of theologians, over questions that theologians are certainly no more qualified to answer than scientists themselves?

Why does Dawkins dress up his opinion as a question? It is a rhetorical device to improve the persuasiveness of his argument. Although I do tend with his sentiment.

P.164 This is actually a quote from Richard Lewontin, but I thought it worth mentioning as it seems quite significantly wrong, and it surprises me that Dawkins includes it:“That is the one point which I think all evolutionist are agreed upon, that it is virtually impossible to do a better job than an organism is doing in its own environment”.

The reasons I disagree with this statement are two-fold:
  • Firstly it seem to me logical that a species would evolve to do a perfect job (impossible to do better = perfect) over an infinite time-frame, but obviously each species has not been in its own environment for infinity, and hence will not be perfectly evolved for it (hence can do better)
  • Secondly, which somewhat contracts the first point above, evolution may be trapped by local maxima and therefore not achieve the global maximum (there are some sufficiently dynamic environments where over an infinite timescale local maxima will be escaped, but there are probably also environments so stable that this would not be the case
But then I’m not an evolutionist sowhat do I know.