GMOs can be safe, they can be unsafe. For instance, if you inserted a gene into a potato to produce poisonous cyanide compounds that would be unsafe. Engineering a rice variety to produce Vitamin A precursors, however, is not likely to be unsafe on a basic scientific level and once the crop has been tested and retested for toxicity and other concerns it is simply not plausible that it would be dangerous.

Other intentions of GMOs can be, for instance, to produce an organism which is naturally resistant to pests or which is better able to grow in a variety of climates. With regards to pest resistance the toxicity of the compound would, of course, have to be tested first - but artificial pesticides are regularly sprayed on crops anyway so I do not see how it is different.

With regards to increased proliferation or better adaptation to more climates, the fear that they will somehow destroy the ecosystem is quite unfounded - if a plant which was slightly better adapted could do this evolution would have assured mutual destruction by know already. Remember that modern varieties of crop plants are probably not well suited to growth outside of human cultivation, being bread to human needs which may be maladaptive in the wild. GMOs are no more capable of taking over the world than their wild type counterparts I assure you - we are years and years away from being able to design something so effectively!

Look at what the basic science of genetic modification is - inserting a known gene or known series of genes into an organisms code under the control of specific gene switches in order to get a desired phenotype. Its not that scary really, an additional gene or an additional few genes added to what is no doubt already thousands. It is merely an accelerated and highly specific form of artificial selection which has been going on for centuries. The exact function of the genes being inserted will no doubt have been ascertained precisely before getting anywhere near the stage of producing a GMO. The only potential risk I can think of is unexpected interactions between metabolites in other biochemical pathways - but this would be picked up on pre release testing and possibly even in silico modelling, and thus can be easily managed.

The fear over GMOs is quite out of control, for instance I have in the past worked with GMOs in a lab setting. This were genetically modified fruit flies (which had the reporter gene Lac-Z added next to genes of interest in order to permit visualisation of gene expression) and Chinese hamster ovary cells which had the cannabinoid receptor artificially added to their repertoire. With both the control were laughably excessive, the fruit flies are no different than their wild type counterparts in terms of any threat of danger and the CHO cells like to die even when good care is taken of them in highly controlled conditions (trust me, they wouldn't last a day if they got out of the lab). Yet still, we had to put up with the highest level restrictions and procedures when working with these in order to absolutely rule out environmental contamination - including intrusive inspections by government staff.