Any organism will adapt to its environment through evolution. Evolution requires that the change could has a selective advantage. In the case of farmed plants we have removed plants from this situation so it is highly unlikely that the GM plant would evolve into some super planet trasher, especially without us noticing. Other things adapting to the GM plant are a different matter, pathogen resistance can be overcome as resistant pathogens will have a selective advantage. But this is also true for conventionally bred resistances.
Plant resistance to pathogens largely relies on a few families of resistance genes that act as antennas for specific pathogen “signatures”. When they detect a pathogen signature they initiate the defence responses required. Given the number of pathogens you can imagine that the repertoire of these resistance genes is quite big. In wild species there are many variants of these genes and the genetic diversity in each population provide for selection, diversification, maintenance and transfer of them, allowing for adaptation to pathogens (arms race idea). During domestication we have inadvertently stopped this flow and fixed a small number of the resitance genes available in the wild in our crop populations. This prevents conventional breeding from being able to adapt to emerging pathogens in the same way that wild populations do as they can’t access the resistance genes because they aren’t present in the breeding population.
In many GM strategies to make pathogen resistant plants the aim is to identify resistance genes from wild populations and move them into cultivated varieties. So, say, moving a wild species potato gene into a domesticated potato without disrupting all of the other benefits breeding has provided such as yield, growth rates, etc. over the wild species. In essence you could say that GM of this type is trying to reverse the negative effects of past selective breeding to increase the adaptive variation available to stop pathogens.
On top of this many strategies are also trying to stack resistances up (introduce multiple pathways of recognising a pathogen) to prevent pathogens overcoming the resistances (similar to multiple complementary antibiotics courses).
I find it unlikely that a superpest would emerge as a result, certainly no more than is observed already with pesticide spraying use. In fact, GM may reduce the risk as you wouldn’t risk spraying off-target organisms which could develop resitance and may become a peat. With GM you would only be targeting the specific organism trying to attack the plant, everything else would be left alone. You could also have two or more variants with resistance to the same pathogen and alternate their use. this would alter the direction of the selective pressure and further reduce the chance of resistance emerging. This is of course theoretically possible with conventional breeding but because of the reasons outlined above and the time it takes it has never been achieved to my knowledge.
Right, dinner break over, back to work.
