Speaking as a singer, the machine orchestra needs to track and anticipate solutions to the singers variation of rhythm, pitch and dynamics. The singer is carefully monitoring the audience and deciding when a strict rhythm is needed to set up their expectations so that he can express by slowing down or speeding up their rhythms or using slightly blue notes or by changing the level of sound or even the nature of the phoneme that he is yelling. He is actually using the audience's pattern judgement to draw them into his performance. Normally when singing with an accompanist it is really a duet, as the singer is depending on the accompanist to read him and anticipate what he is trying to do. The conductor tracks him and will communicate his changes of tempo and dynamic to the orchestra with judgements on how to complete the ensemble. Karaoke is not very pleasing as it is like being frog marched through the music because there is no non verbal consensus on how the music will develop. So if i am working for an audience who has consented to listen to me, i want some sort of intelligence behind me tracking my decisions and forming a reasonable background for them which also has at least a fractal ability to delight the audience.
Speaking as a journeyman pipe organ builder, the onset of new technologies tend to speciate methods and instruments of music. If you add a keyboard to a harp, you are no longer hands on and you have a harpsichord. What you lose in dynamics you gain in control of damping. In pipe organs we have seen electrification of actions, electronic tone generation, sampling, midi control of the instrument and these have all spawned different methods of play and different instruments. You have the electric action pipe organ, the jazz organ, samples you can download and play on your home computer and synthesizers which are a different instrument from where you started. And yet there are still a handful of pipe organ shops in the United States that build tracker (mechanical) organs which could even be gotten with bellows if you want them. This is because that form of the instrument is still considered useful because of the articulation it allows to the musicians for their music.
In the article, there seemed to be something rather naive about the way they were bandying about a few speakers here and there. Speakers are directional and it takes a good sound engineer to design the acoustics in a given auditorium and tune it even before the musicians show up. It takes sufficient drivers to move the air in the room without blowing peoples ears out. Speaker stacks for surround sound can be expensive and for a grand opera are not trivial and you are not talking yet about the amplifiers and generators and boards to control them and then the computer and programming to produce the sound. If you have been to a large music gathering you have probably noticed the huge speaker stacks that enable them to accelerate the air around you without hurting your ears.
I would say the way of singing for Wagner is optimized for an acoustic presentation in a large venue. It is very taxing on the voice to sing that way and takes a lot of athletic ability on the part of the singer to sustain it without electronic amplification although they are usually assisted by the acoustics of the room.
If you mike them, you are no longer doing opera and you will attract a different type of singer and your music will become a different species of musical experience. So the singer will have to decide what direction he would like to follow with his music and what sort of audience he wants to cater to. I would suggest that there is room for all of them in music.