率直な問い
The same qualities that made CRISPR spread through biology — cheap, fast, easy to learn — apply equally to anyone with bad intentions. This is the dual-use problem, and it is genuinely unavoidable: you cannot make a technique easy for researchers and hard for everyone else. The realistic responses are about controlling access to materials, screening synthesis orders, and international norms rather than about restricting knowledge.
Dual-use research of concern in genome editing spans enhancement of pathogen transmissibility or virulence, synthesis of dangerous agents from published sequence, and gene drives as ecological weapons. Countermeasures include DNA synthesis screening by commercial providers, institutional biosafety review, export controls on specified agents, and the Biological Weapons Convention — which lacks a verification mechanism. AI-assisted protein and sequence design has intensified attention on synthesis screening as the practical control point.
The arguments
- Open publication accelerates defensive work — vaccines, diagnostics, therapeutics — more than it aids attack.
- Restricting publication drives work into less visible settings without preventing it.
- Practical controls at the point of materials and synthesis are more effective than controls on knowledge.
- Most serious biological risk remains natural in origin, and preparedness investment addresses both.
- Some results genuinely lower the barrier to causing harm, and publishing them is a choice.
- Synthesis screening is voluntary in many jurisdictions and unevenly applied.
- The Biological Weapons Convention has no verification mechanism, so compliance is unverifiable.
- Cost and skill barriers have fallen substantially and continue to fall.
Where the science actually stands
Dual-use risk is real and recognised across the field. Effective control points are widely agreed to be access to materials and commercial synthesis screening rather than restriction of knowledge, which is both hard to enforce and damaging to defensive work. Governance is uneven internationally.
Unresolved questions
- How to make synthesis screening universal and verifiable.
- How to review results whose publication itself carries risk, without creating unaccountable gatekeeping.
- How to strengthen international verification when the treaty has none.
Sources
- National Research Council · 2004
Biotechnology research in an age of terrorism ↗