Future societies as information economies

Assuming that humanity makes it through the singularity, what will people value and want in the future? What kind of recreational activities will we spend our resources on and get enjoyment from, once we are technologically mature?

Consider housing and in particular home furnishing, one of the activities that humans today spend a non-trivial amount of resources on. Even assuming our future descendants will have a physical form, this is not at all a resource-hungry activity: with fully developed nanotechnology (or even just high-quality 3D-printing and robotics), furnishing a present-day house comes down to specifying what you want and the cost of raw materials. People might still do it, the way people today might configure their computer interfaces to be more aesthetically pleasing, but this hardly feels like where the action is.

In general, any physical goods encounter the issue of replication: once you have one instance of something, what’s stopping a superintelligence with nanotechnology and raw materials from taking it apart and creating two copies? So what will people spend their resources on in a world where the problem of operating with “mere matter” has been solved, and where anything you can observe you can typically also replicate at low cost? When you drop those constraints that currently shape society, what is left?

I argue that, at least, information is, and viewing future societies as information economies is a useful lens.

Consider again the example of movies. It will likely continue to be the case that by using more resources one can produce better movies, even in a technologically mature society, making the movie business possible in principle. In practice, a concern is ensuring that viewers pay for the movie. Theatres already have a norm against photography, but things get complicated by the fact that future people could be cognitively very non-human – in particular, they might have perfect memory and thus reconstruct and distribute the movie once they leave the theatre – and so there is no clear boundary between “consuming” and “recording” media: you might as well allow people to stream the movie in their home theatres.

One possible solution for avoiding information spreading is insurance: an insurance company will reimburse the producers for leaks. In turn, anyone who wants to see a movie needs to have bought insurance whose price is proportional to the risk of them distributing the movie. Importantly, in the future, insurance companies could be much more accurate in their risk estimates, due to the technology allowing privacy-preserving audits as follows:

1. The company sends an AI auditor to go through any information pertaining to you which you permit access to (which could include very personal data, for example your current full brain state).
2. The AI analyses the data and produces a risk assessment.
3. The AI sends the assessment back to the insurance company, after which it self-destructs.

Of course, one would naturally combine this approach with other methods (such as watermarking, monitoring illegitimate distribution and norms against piracy). And it’s unclear whether insurance can be reasonably priced while compensating risk for movies watched at global scale. But privacy-preserving audits extend far beyond movies for enabling trade and coordination, and are potentially key technology for an information-centered society.

The transition to an information economy has of course been happening for a while: with movies in particular, piracy is much more of a concern now than it was in the 1950s. More broadly, intellectual property was a minor issue two centuries ago – you can spy on rival agricultural labourers all you want, but the seeds won’t plant themselves. Disruption caused by automated AI labour is hence “just” an extension of a trend where, for a company, losing more of your data means losing more of your competitive edge, and you increasingly are your data.

Importantly, this applies not just to companies: humans, too, are beings of information. If you had all the information pertaining to me – most notably, the information stored in my brain – you could easily replicate me (assuming nanotechnology); my hardware is cheap and not where my moat is. This extends to social relations, friend groups and local cultures as well – if you had extensive information about each member of a culture, you could create a replica of it. Identity comes down to information.

One corollary is that, insofar as people or cultures want to be unique and ward off cultural appropriation (a fairly common human desire), they are incentivised to protect their information, just as companies protect their intellectual property. This might require quite strong intervention, stronger than is currently common.

Another corollary is that, when thinking of what from today will be of value in the future, you should think in terms of information (and especially information that will not be accessible or preserved by default) – since unlike physical goods or products of today’s labour, information can’t be reconstructed. This viewpoint hence directionally recommends ensuring that existing information stays preserved (e.g. by reliably storing one’s childhood photos) and implementing cheap interventions for collecting more information from the present.