I think that the real problem is the false dichotomy between 'all GMOs are bad' and 'all GMOs are good'. There is clearly a middle path - 'some GMOs are safe, and some aren't, and whether release is acceptable depends on the organisms, the genes, the methods used, and the level of testing done'.
I don't think putting herbicide resistance genes into plant species that don't already have them is a very good idea, because it is a short term solution (selective pressure on weeds will eventually create the same thing anyway), has other environmental costs (the evidence last time I checked is that while glyphosate breaks down quickly, if it gets into waterways in the short period before it has broken down, it can cause considerable damage), and even without any form of gene transfer can end up as a weed (the definition of a weed is a plant growing somewhere where it is unwanted - GMO corn in a GMO soy field is a weed).
Even worse applications can be imagined - for example, imagine a cultivar of corn with genes for a drug, where the corn is toxic because of high levels of the drug - would you want that growing in the next field over from where the corn you are going to eat grows?
I think that it is right for reasonable precautionary measures to be in place whenever substantial genetic engineering of plants is going on, whether it is through transgenic methods, through mutagenically exposed plants, or even through selective breeding combined with sequencing, because there are still high risks.
It is the extreme positions of absolute prohibition or absolute acceptance of all GMOs which are irrational.