Each week, In Theory takes on a big idea in the news and explores it from a range of perspectives. This week, we’re talking about human genetic engineering.

George Church is professor of genetics at Harvard Medical School and stakeholder in various human genetics companies, including Veritas, Editas and Intellia. His laboratory was the first to edit DNA with CRISPR in human stem cells.

Major policy decisions are often impacted by gut feelings. One of the more exceptional technologies affected by public emotion has been genetic modification. Genetically modified “golden rice,” for example, emerged in 2002 as a potential global health solution to a vitamin A deficiency that kills a million people each year, but vandalism of field trials has contributed to the delay of rice research, which already faces technical obstacles.

The communication between emotions and facts is even more crucial as we begin to discuss human genetic engineering as a tool for disease treatment. It is urgent that citizens around the world inform themselves and participate in this rapidly moving set of decisions. What we decide will change all our lives, and there are a few important questions to consider:

1. Can embryos consent? There are more than 2,000 gene therapy human clinical trials in progress. Some must happen at early stages of growth — during child development, for example, to cure blindness. But early interventions, such as with fetal surgeries and medicines, involve risk without consent of the child, so the benefits for the child and society must also be considered. Ironically, doing nothing can entail the biggest risks. For example, delaying retinal gene therapy can result in children who can see light patterns but never learn to interpret faces.

2. Will genetic engineering permanently change our society? Gene therapies are typically done to cure individuals, not alter subsequent generations. But for each such therapy, if effective, the genetic changes will nevertheless be present in future children due to widespread use. This is cultural, not genetic, inheritance — as occurs with other compelling technological innovations, such as cars and phones. Indeed, “cultural inheritance” can spread far more rapidly, in a matter of months rather than over a period of 25 years in generational time. Gene therapies developed to prevent cognitive decline at age 60, for example, could be found to enhance certain skills on Wall Street at age 20 and could enter common use within years.

3. Where do we draw the line? Going forward, developing ethical and policy boundaries on genetic engineering will demand nuance. The lines around our current genetic engineering practices should not be based on categories of technology (genetic therapy vs. genetic counseling, for example), because most categories can contain both good and bad outcomes. Instead, the lines should be drawn based on the outcomes themselves and safety and efficacy, as they are for all new technologies. We should also design our policies to take into account more factors than just the severity of the disorder a genetic treatment attempts to address. If our medical industry approved treatments based on severity alone, we would be missing out on a variety of valuable treatments for seemingly simple maladies such as headaches, sleepiness, insomnia, acne and allergies.

4. Can we reduce risk to embryos? Parents who carry the genetic variation for certain serious inherited diseases (like Tay Sachs) but are themselves unaffected are currently able to avoid passing on the disease to their children by using prenatal testing during pregnancy (or more rarely in vitro “pre-implantation”). This, however, risks the abortion of embryos inherent in the choosing process, which is unacceptable to many parents. An alternative could be to edit sperm-producing cells in the father so that only non-carrier sperm are produced, eliminating the possibility of the disease showing up in embryos. Half of the children will still be carriers – due to the genes being passed on from the mother, so this would not affect the gene pool.

5. Will there be genetic consequences further down the line? The error rate of well-designed CRISPR technology is much lower than normal mutation rates in unmodified DNA, and inspecting modified sperm stem cells can reduce this rate an additional million-fold. But even if genetic effects are initially undetectable, environmental and/or social effects may become evident after many generations. We already monitor many modern discoveries for long-term effects, and tools such as CRISPR should not be an exception. As soon as we discover harm that outweighs benefit, we try to fix it. Yet cars still kill millions of people worldwide each year. Cigarettes were initially approved by health experts and sold in the trillions between 1930 and 2000, ditto for BPA from 1957 to 2008 and DDT from 1940 to 1962. More recently, the painkiller Vioxx was banned after it was shown to cause heart attacks and strokes — but only after 80 million people used it.

6. Will it exacerbate wealth inequality? Commentators have long suggested that genetic engineering will be available only to those who can afford it, resulting in wealth disparities and allowing its advantages to accrue only to a small percentage of society. This risk can be reduced by distributing the high cost of the technology’s development throughout society, as the Orphan Drug Act of 1983, passed to incentivize the development of drugs for rare diseases, meant to accomplish. The cost may also fall on its own, as has happened with electronic technology and genome sequencing. It’s also noteworthy that CRISPR and gene drive technology may provide a path to eliminating malaria or nematode diseases — saving millions of lives per year and breaking cycles of disease, poverty and illiteracy.

7. Are some traits just too complex? Not necessarily. Usually even when studies of complex genetic traits do not reveal single gene variants with large impacts in the body, they can be found via rare natural variants or synthetic biology. Indeed, the most successful therapies target a single gene product in the midst of a complex set of genetic and environmental factors. For example, while hundreds of common genetic variants add tiny effects on height, one gene product (growth hormone) outweighs all of those and is used medically. Similarly, single gene variants have been shown to significantly enhance cognitive tasks in animals. These are being tested for Alzheimer’s disease therapies. As with other therapies, side effects — if any — can be explored and fixed.

8. Are we seeking the ideal human? If so, the idea makes as little sense as seeking out “the ideal transportation”: how do we decide among jets, cars and oil tankers? Lessons of history and evolution show we need diversity, and not just in skin melanin densities. We need immunological, metabolic, cultural and mental diversity — including people with high-functioning autism, obsessive-compulsive disorder, attention deficit hyperactivity disorder, bipolar disorder, dyslexia and narcolepsy. I have mild forms of some of these “disorders” and feel we are not ready to lose them. If you’re in need of convincing, just Google each of those six pejorative terms along with “successful” or “famous.” Indeed, the Israeli military elite unit 9900 embraces autistic skills at interpreting aerial photographs. To prevent such loss, and to adapt to the world of the future, we need public support for such neural diversity. Physics and chemistry have led to leaps in the diversity of our cultural artifacts. Genetics can too.

Explore these other perspectives: