No, robotics and automation is not a contradiction.
The loom and the steam machine represent the mechanisation of production processes. Later came the assembly line and the electrification of the production processes. The next stage is the robotization of the processes (and I inlcude here the numerical controlled machine tools also). What follows, after the "robotization"-era is the current situatuation, sometimes called the IoT era or Industry 4.0.
Each era had their failures and sucesses, but productivity has grown considerably in each era. There is no credible source which disputes the productivity growth. (The argument regarding automation is always related to labour-displacemnet, no credible source argues on the side of overall productivity decrease).
I have to say I do not no anything about the "unimation factory robot" you mention. The other projects seem to be cherry picked and slightly missinterpreted.
The "Halle 54" project was an early project which was a pilot program, probobly something like a prototype program. I think it is normal for early projects to fail to meet all expectations, as I am sure it took a lot of prototyping to get the first loom machine right. There are many other automation projects in the automotive manufacturing industry, which may not be as well known but are increasing productivity. Most automotive manufacturers are steadily increasing the degree of automation of the products. The number of installed and runnig robots is continously increasing since the '90s. My guess it that unexpected things happen when there is a jump (and not gradual growth) in the automation of a manufacturing process (a more recent example is the Tesla production problems). But this is just a guess.
As for the ReThink Robotics story: there are more and more other companies (e.g. Universal Robots, ABB, Kuka) producing so called co-bots (colaborative robots) and the number of installed (and running) co-bots is continously gorwing, therefor the failure of one manufacturer on a steadily growing market is not relevant. The same goes for service robots, if someone considers Baxter/Sawyer a service robot and not a co-bot.
You can take a look at the share prices and market capitalization of the large robotics/automation/machine building companies. Most of them show a growing trend.
Based on the comments below, I would like to re-iterate the conclusions:
There is no credible source disputing the productivity growth due to the use of robots
the number of installed robots doubled in the last 10 years (from approx. 1 million units to more then 2 million units, source is the International Federation of Robotics), proving that the industry uses more and more robots
the projects listed in the question are cherry picked and do not reflect the state of the manufacturing industry.
despite the partial failures of Halle54, Volkswagen is continuously increasing the number of robots manufacturing their cars
My guess is that many projects where a jump in the automation degree of the process instead of a gradual increasted is planned have an increased risk of failiure when compared to project whith a goal of gradualy increasing automation degree. Both cases increase the number of robots in the process, the diffrence is introducing 10 robots at a time for one subprocess or 100 for 10 subprocesses.
Currently in the manufacturing industry, automation in most cases implies the use of robots.