For all the eccentricity of the art, choosing iPods as the delivery device is a classic example of a sound enterprise IT decision. Walsh may have made his money through "gambling", but not in the sense most people understand it.
You have to consider where Android was at 6 years ago when this project commenced. Even assuming a platform switch a year before opening were realistic, find me an iPod sized Android tablet that was on the market in early 2010. What were the specs and price point of that device, assuming it existed? Sitting in 2010 and using your crystal ball, what would you think the likelihood is that you could buy the same device (or another device with hardware specs, drivers, interfaces, and OS that would reliably run your custom app and hardware) in 3/6/12/24 months time? Consider the complexity of supporting multiple different OS and app versions, quite apart from whether it still fits in your custom charging station. Remember the furore from Apple changing the iPod connector this year after it stayed the same for 8 years? Until not so long ago, many vendors changed the connector design with every iteration, and even since standardising on micro USB you're still lucky if the port is in the same place on the next device version).
Having been to MONA, the devices aren't flawless - as another poster mentioned, you often have to select from a list of numerous artworks to find the one you're looking at. This would be a limitation of the number of active RFID devices in the walls as much as in the device itself. I also found them a bit of a distraction from the exhibits.