I know a bit on automation, work R&D in a factory. There's no doubt that it's cheaper in the long run, even with costs related to programming, testing, maintenance, and tweaking. But rushing untested robotics into production will almost certainly increase the costs above manual labor. It's better to maintain the current line while you test out the automation, gradually train workers to monitor systems, and switch over one line at a time. This particular story (painting dots) seems like a great case for automation, so I'd suspect they're working on it.
Also keep in mind that sometimes, manual labor trumps robotics. Let's say there will be 70 million of these backplates shipped, and then no more will be needed. That's a lot of automation investment in a solution for a temporary product. Workers can move from one task to the next with ease, while robotics require a lot of development, logistics, and testing. For a company like Foxconn that will always be manufacturing different things, the investment in human labor might allow for more "pivoting" than robotics. Comes down to what makes the most business sense for them.
Also keep in mind that sometimes, manual labor trumps robotics. Let's say there will be 70 million of these backplates shipped, and then no more will be needed. That's a lot of automation investment in a solution for a temporary product. Workers can move from one task to the next with ease, while robotics require a lot of development, logistics, and testing. For a company like Foxconn that will always be manufacturing different things, the investment in human labor might allow for more "pivoting" than robotics. Comes down to what makes the most business sense for them.