Brain size, social comparison and co-operative machines

A study shows how social comparison is linked to co-operation, human brain size and future developments in machine intelligence 

As our new technologies involve us in ever more rich and more complex sets of relationships, scientists are looking to evolutionary psychology and anthropology to help understand how our capacity for social engagement actually develops, and how this can impact the next generation of interconnected networks, such as those used for mobiles and driverless cars.

One recent paper led by a team of computer scientists at Cardiff University showed that our decision to co-operate with others is based on social comparison; our judgment of their relative social standing. 

Lead author of the study, Professor Roger Whitaker, from Cardiff University’s School of Computer Science and Informatics, says that, ‘we have shown that over time, evolution favours strategies to help those who are at least as successful as themselves.’ This cooperation is essential for personal and social prosperity, and has been suggested as being a driving factor in our evolution as a species. The new study suggests that this understanding can also influence the cooperation of interconnected devices.

Researchers collaborate to research collaborations

The current paper on co-operation is itself the product of co-operation between diverse and successful researchers, including co-author Professor Robin Dunbar, head of the Social and Evolutionary Neuroscience Research Group at the University of Oxford.

Professor Whitaker explains how this came to be: ‘It was around 2009-2010 when I first bumped into Robin Dunbar the collaborator on this paper. We had a joint grant supported by the European Commission for a project around social networking, looking at it from a number of different disciplinary approaches, from the human being right through to theoretical physics. Part of the inspiration for this work came from our collaboration on that project.’ 

This issue of co-operation arose through collaboration with another set of engineers at Queens University Belfast around the next mobile telephony communications - 5G, which will be faster. Currently mobile phones work with the technology of base stations, but though the industry specifications for 5G haven’t been agreed, one of the possibilities is looking at device-to-device communication. ‘I can use my phone to hop on to another network through somebody else's phone, there is a relaying function that phones may also play for other people,’ says Professor Whitaker. 

The project exploring this is called GigaMobile, sponsored by the EPSRC (Engineering and Physical Sciences Research Council) and involved industry partners such as Samsung and Telefonica. Which is where the issue of co-operation begins to emerge, ’as soon as you start talking about piggybacking or using someone else's device to connect to the network, you soon come across a problem of co-operation,’ says Professor Whitaker. ‘Why on earth should I let someone else use my resources my battery to do that? You need some kind of incentive system or reward or payments system. Setting up a payment system at a micro-level like that is such potential overhead.’ 

Social Comparison 

They received funding to look at this telecommunications technology problem from a co-operation point of view, and this produced the results of the study. ‘If two devices calm into a range of each other what can they do? The first thing they can do is compare themselves, this lead us into the literature around social comparison and collaboration with Robin Dunbar.’

Evolutionary psychology and anthropology became a framework for addressing an engineering problem. ‘We've had to just ignore disciplines and look for solutions,’ he says.  What encourages and incentivises co-operation?

They used computer modelling, simulated representations of human beings to run hundreds of thousands of ‘donation games’ to understand decision-making strategies, and why certain kinds of co-operative behaviour strengthens over time. A decision to donate in the game incurred a cost to the giver while the receiver benefited – then social ‘reputation’ was updated. 

‘This is highly relevant to humans in a number of ways,’ says Professor Whitaker. ‘This species that we’re in is really pro-social, we do things on a voluntary basis, we make it donations almost intuitively. It's kind of a hallmark of what we do - whether it's letting someone out in traffic, or giving someone a tip, we’re always doing this as human beings. We operate in groups and we are able to size up other people, implicitly we are used to making judgments about of others.’

In this way the research connects cooperation in groups and the practice of comparison. It also provides support for the Social Brain Hypothesis, the idea that humans have a large brain because they have evolved in complex social groups, first proposed by Professor Dunbar. ‘The idea that we may well have extended huge amounts of cognitive energy on sizing each other for purposes of deciding who we are going to help or not, that's something which carries forward – those that are good at that have a propensity to carry forward into the next generation.’ explains Whitaker.

From an engineering point of view, Professor Whitaker believes that the research gives us a basis, ‘to think about some of these autonomous systems and intelligent systems there are now starting to emerge. This is the first time in history that we've been in the position where these systems are emerging in a very complex way.’ Whether it is 5G phones or driverless cars, machines will need to be engineered to co-operate. ‘We think this is the beginning of being able to make technology or artificial intelligence and autonomous systems a bit more human.’