From the perspective of a certified gene jockey (albeit who doesn't work on plants), conventional transgenesis is risky. Gene insertion can occur randomly, and even though most of the genome was historically considered to be "filler", it turns out that there are a wide variety of non-coding RNAs in those empty regions. If you insert a gene in the wrong spot, you can inadvertently disrupt an endogenous gene - to unknown effect. That's why it's so important to map your inserts, particular for a long term project like this. Moreover, conventional gene insertions are often comprised of repeats that contain multiple copies of the promoter::coding region modules. Their repetitive nature makes them increasingly subject to gene silencing over time. In essence, they are treated as foreign invaders, like retroviruses, protecting the integrity of the genome.
Given the advent of CRISPR/Cas9 gene editing technology and our current ability to insert a single copy of a gene at precisely defined positions in the genome, I fully expect that this approach will be replicated with greater success as we move forward. I don't think that it makes sense to introduce a strain with known weaknesses into the wild though.