Commercial genetic testing is racing ahead—will it change who we are, even if it doesn’t work?

Disclaimer: This article is for information only and is not medical or legal advice.

Background

Over the past 15 years, genetic testing has gone from a niche clinical service to a mass-market product. A saliva kit can estimate your heritage, flag disease risks, tell you how you metabolize caffeine, or suggest which vitamins to take. In fertility clinics, increasingly sophisticated add‑ons promise to screen embryos for chromosomal problems, single‑gene disorders, and even lower risk for common diseases through polygenic scores. In obstetrics, noninvasive prenatal testing (NIPT) uses fragments of fetal DNA in maternal blood to assess chromosomal abnormalities early in pregnancy.

What began as an empowering story—more information for individuals—has now become entangled with thornier questions. The science behind many offerings is uneven; predictions for complex traits often depend strongly on ancestry, environment, and statistical happenstance. Even so, the tests are changing behavior: couples decide which embryos to implant, people alter health plans, and families revise their sense of identity after an unexpected ancestry result. A new book surveyed by Ars Technica argues that these social effects could accumulate, reshaping human diversity and social dynamics, even when the tests underdeliver scientifically.

The claim sounds paradoxical: how can tools that don’t work well still have outsized consequences? Consider three forces that operate regardless of predictive accuracy:

• Market incentives reward bold claims and differentiation, not caution.
• Behavioral responses follow perceived risk rather than calibrated probabilities.
• Inequalities in access turn small average effects into large social gradients.

Put together, this cocktail can change who gets born, how we define groups, and which genomes are more likely to persist—not by rewriting DNA, but by nudging choices.

What happened

Ars Technica reviewed a new book that interrogates the rise of commercial genetic testing—particularly polygenic risk scores (PRS) for health and embryo selection—and argues that we may be transforming the genetic and cultural landscape without the guardrails of clear evidence or strong regulation.

The core concerns highlighted by the book, and echoed by many geneticists and ethicists, include:

1. Predictive limits of polygenic scores

   • For common diseases and traits (heart disease, type 2 diabetes, height, educational attainment), risk is influenced by thousands of genetic variants plus environment. Polygenic scores trained mostly on European-ancestry datasets perform less well when applied to people with different ancestries. In independent studies, performance drops by a factor of two to five outside the discovery population. That means the very groups long under‑represented in genetics get the least reliable predictions.
   • Even within matched ancestries, variance explained is modest for most outcomes, and scores are sensitive to study design choices, subtle population structure, and unmeasured environmental confounders.

2. Embryo selection’s small absolute gains

   • Embryo choice using PRS (sometimes marketed as PGT‑P) aims to pick the embryo with the lowest predicted risk among siblings created via IVF. But IVF typically yields only a handful of embryos. With small sample sizes, the best available PRS offers only marginal reductions in absolute disease risk or minor shifts in continuous traits. Gains measured in population averages do not translate cleanly to family-level guarantees.
   • Pleiotropy—the fact that one genetic variant can influence multiple traits—creates trade‑offs. Lowering risk for one outcome might nudge risk for another in the wrong direction. For example, variants associated with lower cardiometabolic risk may intersect with growth or reproductive traits, while height-boosting alleles connect to elevated cancer risks. Selecting on composite scores can entrench opaque value judgments about which risks “count.”

3. Prenatal screening’s expanding scope and uneven accuracy

   • NIPT is robust for common trisomies, but as companies extend panels to rare microdeletions and single‑gene conditions, positive predictive values can plummet in low-prevalence settings. False positives (and false reassurance) can drive irreversible choices in the fog of anxiety.
   • The mere availability of a result—even with caveats—can steer decisions. In some countries, widespread screening has dramatically reduced births with certain chromosomal conditions, rekindling debates about modern eugenics and disability rights.

4. Identity, kinship, and the ancestry marketplace

   • Direct-to-consumer ancestry testing has produced seismic shocks in family narratives: unexpected parentage, undisclosed adoptions, and newly discovered relatives. Meanwhile, estimates of heritage percentages turn on statistical reference panels and algorithmic thresholds. The numbers feel precise but are interpretive.
   • These shifting identities ripple into social and political domains—eligibility for tribal affiliation, claims to citizenship, and cultural belonging—despite the tests’ limited capacity to define identity.

5. A regulatory patchwork and powerful incentives

   • In the United States, federal oversight of laboratory-developed tests is fragmented; professional societies have advised caution on embryo polygenic testing, yet clinics market add-ons to differentiate services. In Europe, new in vitro diagnostic rules increase scrutiny, but enforcement varies by country. Globally, consumers often can buy testing that outpaces clinical consensus.
   • Data brokerage and research monetization mean a consumer’s DNA can be repurposed for studies, targeted advertising, and even law enforcement lookups through genealogy databases—outcomes not always fully understood at the point of consent.

The book’s provocation is not that genetics is bunk. Clinical genetics is transformative when variants have large effects, when there is a clear causal mechanism, and when test performance is validated in the population being served. Rather, the claim is that commercial genetics frequently moves the goalposts: it thrives in gray zones where weak signals can still shape powerful stories—and thus behaviors.

Key takeaways

• Weak predictions can make strong markets. If people act on uncalibrated risk estimates, the social and demographic effects can be large even when the underlying science is modest.
• Embryo selection for complex traits is statistically constrained. With few embryos and many trade‑offs, average gains are small; yet perceived optimization could still reduce certain alleles’ prevalence and narrow genetic diversity over time in subpopulations with high uptake.
• Portability gaps amplify inequities. Polygenic tools trained predominantly on European cohorts are less accurate for other ancestries, potentially steering non‑European families toward worse or misleading decisions.
• Screening is not neutral. The choice of what to screen for expresses social values; overreach in prenatal or embryo testing can implicitly deem some lives less worth living, raising ethical concerns voiced by disability communities.
• Data practices outlive the test. Once collected, genomic data can be used and re-used in ways that reshape privacy, law enforcement, and insurance landscapes.
• Regulation lags incentives. As long as oversight is light and demand is intense, companies will market “informational” tools whose risk framing may exceed what evidence supports.

How the science and the sociology collide

The illusion of fine control in complex traits

Complex traits are governed by innumerable small genetic effects that interact with each other and with environment. Polygenic scores summarize these signals into a single number—useful for research and sometimes for stratifying risk in large populations. But the seductive precision of a percentile ranking can hide key caveats:

• Environment matters deeply. A high genetic risk for type 2 diabetes may never manifest with healthy diet, exercise, and medical care; a low score is not a shield against an obesogenic environment.
• Genetic architecture varies by ancestry. If the training data underrepresent your background, the score may over or under-estimate your risk.
• Measurement error and selection bias. Many GWAS rely on volunteer biobanks with healthier, wealthier participants; the resulting models carry those biases forward.

When such a score is used to choose between embryos, you compound the uncertainties: siblings share most of their genome, so the spread in predicted risk is small; any antagonistic pleiotropy or noise looms large relative to the expected benefit.

The behavioral economics of genomic risk

People do not respond to risk numbers the way statisticians wish they would. They overweight rare catastrophic outcomes, seek reassurance in “low risk” labels, and anchor on stories. In reproductive contexts—where decisions are irreversible—the pressure to act on any available signal is intense. That is why even low-accuracy tests can have high-impact consequences. Consider:

• A false positive on a rare microdeletion panel can lead to invasive follow‑up tests or pregnancy termination.
• An embryo score that promises “lower lifetime risk” could steer implantation choices in ways that reduce the transmission of alleles with uncertain or even beneficial effects in other contexts.
• An ancestry readout can reshape identity, family relationships, and legal standing, despite being a statistical inference with confidence intervals rarely shown in consumer apps.

Inequality and the possibility of stratified biology

If only affluent families can afford IVF with polygenic screening, small effect sizes could still create stratification over time. The first generation might see negligible differences; the tenth could show discernible shifts in polygenic profiles for certain traits among the socioeconomically advantaged. Meanwhile, underperformance of the same tools in underrepresented populations could deepen disparities through misclassification or missed benefits.

Even without embryo selection, the clustering of health insights among the wealthy can translate into earlier detection and better prevention—biological advantages layered on top of social ones.

What to watch next

• Professional guidance solidifying on embryo polygenic testing. Multiple societies have advised against routine use of PGT‑P for complex disease risk, citing insufficient evidence and ethical concerns. Watch for updated statements as new validation data arrive—or fail to.
• Regulatory moves on laboratory-developed tests (LDTs). In the US, proposals to expand FDA oversight of LDTs could touch both DTC health reports and clinic-based embryo screening. In the EU, the In Vitro Diagnostic Regulation (IVDR) is still phasing in; enforcement will shape what reaches patients.
• Litigation and consumer protection. Expect more actions over misleading claims, undisclosed data reuse, or harms from false positives. Clear precedent could recalibrate marketing language across the industry.
• Diversity in genomics research. Efforts to recruit participants from underrepresented ancestries and to develop ancestry-aware models will determine whether polygenic tools can become equitable and clinically meaningful.
• Privacy and policing. Law enforcement’s use of genealogical databases will continue to test social norms around consent and familial privacy. Future cases could spur opt‑out defaults or statutory guardrails.
• Cultural feedback loops. As more individuals confront surprise ancestry or health findings, norms around parentage disclosure, donor anonymity, and identity claims will evolve—likely faster than scientific accuracy does.

Practical advice for consumers

• Treat complex-trait results as conversation starters, not verdicts. Ask: What is the absolute risk? How does this apply to my ancestry and family history? What’s the positive predictive value in my setting?
• Seek confirmatory testing for medical decisions. Use clinically validated tests under medical supervision before changing medications, undergoing procedures, or making reproductive choices.
• Read consent and privacy policies. Understand data retention, sharing with third parties, and options to delete your data.
• Be skeptical of “optimization” claims in fertility. Ask clinics for peer‑reviewed evidence, expected absolute risk changes, and statements from professional bodies.
• Consider the ethical dimension. Screening choices speak to values. Engage with perspectives from disability advocates and bioethicists.

FAQ

What is a polygenic risk score?

A polygenic risk score (PRS) aggregates the small effects of many genetic variants associated with a trait or disease into a single number meant to summarize relative genetic predisposition. PRS can be useful for research and sometimes for stratifying screening intensity in clinical settings, but their performance depends on ancestry, environment, and study design.

Does selecting embryos by PRS guarantee a healthier or smarter child?

No. For complex traits, expected gains are small because siblings share most of their genomes and because many variants affect multiple traits with trade‑offs. Environmental factors further dilute predictive power. Professional societies generally advise that embryo selection using PRS should not be routine.

Are direct-to-consumer ancestry percentages precise?

They are statistical estimates based on reference populations. Different companies use different panels and methods, so your percentages can shift over time or vary across providers. They are informative for broad regional ancestry but are not definitive markers of identity or legal status.

How accurate is noninvasive prenatal testing (NIPT)?

NIPT is highly sensitive and specific for common trisomies such as Down syndrome when used in appropriate populations, but its positive predictive value declines for rare conditions. Any positive NIPT result for rare anomalies should be confirmed with diagnostic testing before decisions are made.

Could genetic testing reduce human diversity?

Yes, indirectly. If many people use screening to avoid certain variants—through prenatal testing, carrier screening, or embryo selection—allele frequencies can shift within subpopulations. If uptake is concentrated among wealthier groups, social stratification can map onto subtle genetic differences over generations. The effect’s pace and magnitude depend on participation, trait architecture, and trade‑offs.

What protections exist for my genetic data?

Protections vary by country. In the US, GINA limits genetic discrimination in health insurance and employment but not life, disability, or long‑term care insurance. Company policies, data-sharing agreements, and law enforcement access through genealogy sites are important to review.

The bottom line

Commercial genetics is not a monolith. Some tests bring clear clinical value; others are better considered entertainment or exploratory research. The danger is not that every test is a sham—it’s that market forces and human psychology can turn modest signals into consequential actions. We risk reshaping reproduction, identity, and inequity before we’ve done the slow work of validation, governance, and public deliberation. If the book Ars Technica reviews has a single plea, it is this: move at the speed of evidence, not marketing.

Source & original reading: https://arstechnica.com/science/2026/02/have-we-leapt-into-commercial-genetic-testing-without-understanding-it/