The AI Genetics Revolution Is Coming (www.thenewatlantis.com)

AI is poised to become as transformative for genetics as calculus was for physics: large language models and related machine‑learning tools are beginning to “read” and “write” the language of DNA, turning a previously intractable tangle of multigenic interactions into tractable patterns. Ambitious projects like Evo 2 (trained on ~9 trillion nucleotides across nearly all known species) are already predicting pathogenic mutations, revealing gene networks, and generating novel sequences; other work includes DNA‑to‑face reconstructions, AI‑designed switchable synthetic sequences (Yale/MIT/Harvard), and bespoke CRISPR proteins created by language models to reduce off‑target effects. These advances accelerate both somatic therapies (several FDA‑approved treatments exist, e.g., for sickle cell) and more contentious capabilities such as embryo screening and synthetic embryo production via in vitro gametogenesis (IVG). The implications are profound and immediate: enormous medical upside (precision gene therapy, resilient crops, biomanufacturing) sits beside risks of de facto eugenics—“designer” offspring through overproduction + AI‑powered trait prediction—and widening inequality if enhancements are commodified. Geopolitically, China’s rapid AI progress (e.g., DeepSeek’s R1) and stated biotech ambitions threaten to shift global bioethical norms, shrinking the window for Western policy to shape safe, equitable rules. Biological complexity still poses hurdles, but the policy, ethical, and regulatory debates must accelerate to match the technical pace.