Genetic Engineering in plants

First, the engineered crops themselves could become weeds, a broad term that covers plants with undesirable effects. Second, the crops might serve as conduits through which new genes move to wild plants, which could then become weeds. Third, crops engineered to produce viruses could facilitate the creation of new, more virulent or more widely spread viruses. Fourth, plants engineered to express potentially toxic substances could present risks to other organisms like birds or deer. Fifth, crops may initiate a perturbation that may have effects that ripple through an ecosystem in ways that are difficult to predict. Finally, the crops might threaten centers of crop diversity.

Although few problems of the sorts listed above would be expected to surface within the three-to-four-year time frame that the new crops have been in widespread use, the good news is that there have been no serious environmental impacts.

Pesticide-making plants pose three risks to the environment. Two are old risks, and one is new.
The old risks are the ones seen with overuse or misuse of any pesticide: evolution of pesticide-resistant insects and accidental killing of desirable insects. These risks occur regardless of the source of the pesticide, be it naturally occurring, "organic", synthetic, or genetically engineered. In fact, some insects are resistant to plants' own pesticides, so farmers need to rotate their crops, and try to predict whether resistant bugs will be prevalent in a given year.

The one new risk is the possibility that genetically engineered plants will spread their genes for making pesticides to wild relatives, which may make the first two risks worse.

Keep in mind that all plants, both wild and domestic, have been locked in an evolutionary arms race with insects that try to eat them since insects first emerged. All plants produce insecticides (to a greater or lesser extent). Some insects have evolved to cope with specific insecticides, and eat the plants that make them. People's deliberate use of pesticides has increased the selective pressure on insects to resist them, but this alone doesn't mean that they'll succeed in becoming resistant.

For example, it's likely that some insect species evolved resistance to the bacterial pesticides generally known as Bt a very long time ago. Scientists first noticed this resistance in the early 1970s, long before plants were genetically engineered to produce Bt, and only soon after Bt became popular as an "organic" pesticide. Increasing use of Bt has slowly increased the proportion of insects resistant to it, first with Bt's use in "organic" farming, and now with its use in genetically engineered plants (corn and cotton most notably). If we continue to increase our use of Bt, and don't develop any other pesticides, it's likely that we'll see more and more Bt resistant insects. Conversely, if we maintain a large number of pesticides, and use them wisely, it's possible we may be able to hold down the proportions of insects that resist any or all of them.

In any case, insects are not very likely to become as entirely resistant to as many pesticides as certain bacteria ("superbugs") have. Some bacteria are able to pass around genes in ways other living things don't -- this method is called conjugation. Other creatures must evolve resistance through natural selection, which is far slower, and far less certain to succeed.