It turns out that it can tell you a lot.
This is what I took away from a very interesting talk given by Princeton sociologist, Janet Vertesi, on “the social life of robots.” As it turns out, it is not about the robots but about the people that control them — as opposed to the opposite that we apparently fear nowadays. Vertesi has studied various NASA inter-planetary teams that are comprised of a robot sent out years before and a team telling the robot what data to collect. All the while, it is subject to difficult choices to ensure the robot doesn’t die. And can robots die … well, this xkcd cartoon apparently brings a tear to its team’s eyes.
While the emotional attachment of scientists to robots is potentially interesting, what I want to talk about here was Vertesi’s descriptions of two different teams with robots at two different planets. I’m going to call them Team PowerPoint and Team Excel. (Why? Because Vertesi is still working with them and prefers to keep them somewhat disguised. Vertesi has, however, written a book on the Mars Rover missions so you can get a clue there but in one of the papers that her talk was based on she hides the mission locations.)
As it turns out, that is pretty secondary to the findings. On Team Excel, the robot has a number of instruments but separate teams manage and have property rights over those instruments. The structure is hierarchical and the various assignments the instruments are given are mapped out in Excel. By contrast on Team PowerPoint, no one team owns an instrument. Instead, all decisions regarding, say, where to position the robot are made collectively in a meeting. The meetings are centered around PowerPoint presentations that focus on qualitative trade-offs from making one decision rather than another. Then decisions are taken using a consensus approach — literally checking if everyone is “happy.”
What is fascinating about this is that the type of data collected by each team is very different. On Team Excel where each instrument is controlled and specialised to its task, the data from them is very complete and comprehensive on that specific thing — say, light readings, infrared etc. On Team PowerPoint, there are big data gaps for each instrument but there appear to be more comprehensive deep analyses of particular phenomenon where all of the instruments can oriented towards the measurement of a common thing. This is a classic trade-off between specialised knowledge and deep knowledge. And it is a trade-off that each of these missions needs to make. What is extraordinary is that they bake the trade-off into their organisational structure and also decision-making tools — literally emphasising different apps in Microsoft Office.
There are other interesting elements that come from all this. For instance, different disciplines associated with different teams are more likely to collaborate on scientific publications in Team PowerPoint than Team Excel. Also, there is a different philosophy of commons ownership with the entire robot being owned in Team PowerPoint whereas individual instruments are owned in Team Excel. But the conflicts in Team Excel can arise and are given weight to various bilateral trades and a strange non-sharing norm in their own data economy. By contrast, Team PowerPoint shares everything — even publicly.
From this, it sounds like Team PowerPoint is more functional. And it was was easy to get that impression. But the important take-a-way is that the data being collected by each was so different. We tend to think of space exploration missions as costing lots of money but there not being much after the fact choice of precisely what scientific information will be collected. In this case, the choice of organisational structure, that was made for reasons more random than you’d think, led to very different outcomes. Literally, our knowledge of one planet will be very different from our knowledge of another. Both are very informative but it is somewhat worrying that the type of knowledge was not part of the initial funding decision and that got sorted out later on.