Summary: There are historical precedents where bans or crushing regulations stop the progress of technology in one industry, while progress in the rest of society continues. This is a plausible future for AI. Epistemic Status: My intuition strongly disagrees with other people here. I hope to explain my intuition, and provide enough historical evidence to make this intuition at least plausible.
I disagree with putting atomic gardening on a list of technologies that are underutilized due to societal nervousness. The majority of grapefruits grown in Texas, and rice grown in California, are of varieties produced by atomic gardening (there are other varieties grown around the world). The reason atomic gardening never dominated agronomy is that it is labor intensive. Almost all mutations are useless. If your goal is to improve yield, you might plant a million mutated seeds and still not see a useful phenotype (but if your goal is immunity to a virus, you probably wouldn't have to plant that many seeds). Hybridization and transgenesis allows for copying from useful phenotypes that are already present in nature, so they're easier to find.