Major artificial intelligence (AI) developers, including OpenAI, Anthropic, Microsoft AI, and Google DeepMind, have issued a joint call to the United States Congress to implement mandatory safety screenings for all synthetic DNA purchases. The coalition, working alongside biotechnology experts, argues that existing voluntary practices among manufacturers are insufficient to prevent malicious actors from ordering genetic sequences capable of recreating dangerous pathogens. Synthetic DNA, which is artificially manufactured genetic material used for vaccine development and scientific research, is currently ordered and shipped globally without standardized regulatory oversight. The signatories emphasize that mandatory screening represents one of the most effective and least disruptive methods to enhance biosecurity.

In addition to pre-purchase screening, the proposal advocates for strict, mandatory record-keeping requirements. This would allow for the tracking of suspicious orders and act as a deterrent, ensuring that potentially harmful activities can be traced back to their source even if individual sequences appear benign in isolation. While the risk of weaponized biology is not new, the authors highlight that the rapid advancement of AI models has lowered historical knowledge barriers, noting that current systems can outperform PhD-level virologists regarding complex laboratory procedures. Experts warn that as these systems improve, the technical hurdles that once prevented unauthorized biological weapons production are rapidly eroding.

This concern mirrors international warnings from February, when researchers from institutions such as Johns Hopkins University, the University of Oxford, and Stanford University called for robust safeguards, including watermarking and audit logs, to verify users and track potential misuse of sensitive biological datasets. These researchers noted that while open-access data remains crucial for scientific discovery, a specific subset of new data requires higher security standards.

The situation in Europe is also evolving, as the European Commission introduced the EU Biotech Act in 2025 to address biosecurity risks in biotechnology and biomanufacturing. The proposal identifies synthetic nucleic acid sequences as products of concern, noting that the absence of unified national rules weakens regional prevention efforts. The EU framework, similar to the US proposal, intends to enforce customer verification and the reporting of suspicious orders to ensure a consistent regulatory standard across member states.

Major artificial intelligence (AI) developers, including OpenAI, Anthropic, Microsoft AI, and Google DeepMind, have issued a joint call to the United States Congress to implement mandatory safety screenings for all synthetic DNA purchases. The coalition, working alongside biotechnology experts, argues that existing voluntary practices among manufacturers are insufficient to prevent malicious actors from ordering genetic sequences capable of recreating dangerous pathogens. Synthetic DNA, which is artificially manufactured genetic material used for vaccine development and scientific research, is currently ordered and shipped globally without standardized regulatory oversight. The signatories emphasize that mandatory screening represents one of the most effective and least disruptive methods to enhance biosecurity.

In addition to pre-purchase screening, the proposal advocates for strict, mandatory record-keeping requirements. This would allow for the tracking of suspicious orders and act as a deterrent, ensuring that potentially harmful activities can be traced back to their source even if individual sequences appear benign in isolation. While the risk of weaponized biology is not new, the authors highlight that the rapid advancement of AI models has lowered historical knowledge barriers, noting that current systems can outperform PhD-level virologists regarding complex laboratory procedures. Experts warn that as these systems improve, the technical hurdles that once prevented unauthorized biological weapons production are rapidly eroding.

This concern mirrors international warnings from February, when researchers from institutions such as Johns Hopkins University, the University of Oxford, and Stanford University called for robust safeguards, including watermarking and audit logs, to verify users and track potential misuse of sensitive biological datasets. These researchers noted that while open-access data remains crucial for scientific discovery, a specific subset of new data requires higher security standards.

The situation in Europe is also evolving, as the European Commission introduced the EU Biotech Act in 2025 to address biosecurity risks in biotechnology and biomanufacturing. The proposal identifies synthetic nucleic acid sequences as products of concern, noting that the absence of unified national rules weakens regional prevention efforts. The EU framework, similar to the US proposal, intends to enforce customer verification and the reporting of suspicious orders to ensure a consistent regulatory standard across member states.