> "Now, the interesting question is why, specifically, we would consider that the GMO is riskier than a wild conventional crop"
We are coming to a point in genetic engineering technology where entirely custom organism genomes will be able to be created with four bottles of chemicals: (A)denine, (C)ytosine, (T)hymine and (G)uanine, a computer code specifying the desired sequence, and a computerized melecular assembly machine. Limited examples of this have already been carried out, and its general application is not far off.
The appropriate term then becomes "synthetic biology" not the more limited "genetic modification".
We already see genes from distant species spliced in to other species (fish genes into tomatoes etc).
The answer to the "why more risk" question is that the combinatoric possibilities for novelty of genome and novelty of effect are much greater in genetic engineering than in evolutionarily selected natural mutation.
Ordinary mutation has characteristics like that it is usually only an incremental change (genetic-informationally) from the pre-mutated genome. It is true that even incremental informational change in the genome can lead to large effects in the phenotype (the organism), but with current day and near future genetic engineering, there is no longer a restriction to incremental informational change to the genome.
Most variations will, as usual, not be viable, but if one is by chance or design, it could easily be very different than anything seen in earth life so far, because its synthetic genome can be arbitrarily different.