Two of the most powerful companies in artificial intelligence have done something rare in tech: they asked the government to regulate an industry they do not operate in. OpenAI and Anthropic have co-signed a letter urging US Congress to impose strict new rules on AI synthetic DNA regulation, not to limit their own models, but to stop those models from helping someone synthesize a biological weapon.
The move stands out for a simple reason. Sam Altman and Dario Amodei lead companies that compete fiercely for talent, compute, capital, and customers. Yet they signed the same letter, and that matters because it shows how seriously both companies view the risk. In practice, their agreement is a signal that the threat has moved beyond theory.
Summary
AI leaders urge Congress to regulate synthetic DNA sales
The letter focuses on the synthetic DNA supply chain, the commercial network of vendors that manufacture custom genetic sequences for research, medicine, and industrial biology. Right now, anyone with enough money and a shipping address can order synthesized DNA from a growing list of providers. Meanwhile, the process is becoming more automated and more affordable.
What concerns OpenAI and Anthropic is what happens when powerful language models enter that workflow. AI can help a user understand complex biological processes, troubleshoot synthesis problems, evaluate dangerous pathogens, and identify sequences with specific traits. Because of this, the combination of accessible AI and accessible DNA synthesis creates a pathway that did not meaningfully exist five years ago.
The letter does not ask Congress to restrict AI models themselves. Instead, it targets the physical chokepoint: the moment a synthesis order leaves a vendor’s facility.
Proposed rules for DNA and RNA synthesis vendors
The framework outlined in the letter is specific and operational. It is not a broad call for “more oversight.” Instead, it sets out three concrete requirements for DNA and RNA synthesis vendors.
Mandatory screening against dangerous sequence databases
Every synthesis order would need to be screened against established databases of known dangerous sequences before fulfillment. In effect, vendors would run each requested sequence through a biosecurity filter to check whether it resembles or matches genetic material tied to select agents, toxins, or other biological threats.
Customer verification for DNA synthesis orders
Customer verification would shift from optional to mandatory. At present, synthesis vendors vary widely in how carefully they check who is placing an order. Under the proposed DNA synthesis compliance framework, that verification process would be standardized and required across the industry.
Comprehensive risk assessments before shipping
Before any order ships, vendors would need to conduct comprehensive risk assessments at the order level. That means reviewing not only the sequence itself, but also the context of the request, including who is ordering, the stated purpose, and whether the full set of details raises flags that require extra scrutiny or refusal.
Together, these measures would push the DNA synthesis industry from largely self-regulated practices to a federally mandated compliance system.
Why OpenAI and Anthropic see bioweapon risk in AI
The push for regulation did not come from vague worry. OpenAI has been running internal red-teaming exercises since early 2024 to test whether its large language models could meaningfully assist someone evaluating biological threats. The results were concerning enough to persuade the company that voluntary industry norms were no longer enough.
That point matters because red-teaming is meant to pressure-test a system under adversarial conditions. In other words, it asks what the worst-case misuse might look like. OpenAI’s own evaluations surfaced credible bioweapon risk scenarios, and those findings helped shape the case for regulation.
Dario Amodei has also been outspoken about biological risks from AI. Anthropic’s research has flagged the potential for large language models to provide meaningful “uplift” — a biosecurity term for assistance that helps someone move closer to a dangerous capability they could not reach on their own — to people trying to navigate complex biological processes. As a result, Amodei and Sam Altman co-signing the same document reflects a rare moment of industry consensus.
That cross-competitor alignment sends a message to Congress. When the two leading AI labs agree that a threat is serious enough to require government intervention, the argument for doing nothing gets weaker.
What AI synthetic DNA regulation could mean for biotech markets
If adopted, the proposal could reshape several markets at once, though not in the same way.
For DNA synthesis companies, screening systems, customer verification tools, and order-level risk assessments would bring real costs. Smaller vendors with thin margins could find those compliance requirements especially hard to absorb. As a result, the market could consolidate, with larger firms that already have compliance infrastructure gaining an edge over smaller competitors. New entrants would also face higher barriers.
At the same time, the proposal could create a tailwind for another group of companies. Biosecurity technology firms, especially those building sequence-analysis platforms, biological threat databases, and screening tools, could see stronger demand if federal rules create a formal need for their products. In that sense, mandatory DNA synthesis compliance could open a new market for biosecurity technology investment.
The picture is more complicated for decentralized science. Several blockchain-based DeSci projects sit at the intersection of AI and biological research, offering open platforms for genomic data sharing, drug discovery, and synthetic biology tooling. However, mandatory screening at the DNA synthesis layer could create friction for open-access workflows that depend on easy movement between research steps. Projects built around permissionless access to biological data may need to redesign around compliance checkpoints they were not built to handle.
The broader policy backdrop also matters. This push fits within the wider 2026 AI policy momentum around dual-use technology oversight. Synthetic biology’s overlap with AI makes it one of the clearest examples of dual-use risk: technology that is essential for legitimate science but potentially catastrophic in the wrong hands. In that context, the OpenAI-Anthropic letter is part of a larger debate over where the line should be drawn.
Whether Congress moves quickly, moves slowly, or does not move at all will determine whether the letter becomes a policy milestone or a historical footnote. The screening infrastructure, verification systems, and risk assessment protocols already exist in parts of the industry. The open question is whether they become legally required, universally applied, and consistently enforced.
FAQ
Why are OpenAI and Anthropic asking for regulation of synthetic DNA sales?
Both companies say AI models can meaningfully assist people evaluating and potentially developing biological threats. Combined with the accessibility of commercial DNA synthesis, they argue this creates a serious biosecurity risk that voluntary industry norms cannot handle on their own. Their letter to US Congress calls for federal action on the synthesis supply chain.
What specific regulatory measures are proposed in the letter?
The letter proposes three main requirements: mandatory screening of all DNA and RNA synthesis orders against databases of known dangerous sequences, mandatory customer verification for every synthesis order, and comprehensive risk assessments for each order before it ships.
How has AI demonstrated potential risks in aiding bioweapon development?
OpenAI said it began internal red-teaming exercises in early 2024 to test whether its large language models could meaningfully assist someone assessing biological threats. Those exercises produced results serious enough to push the company toward regulatory action instead of relying only on self-regulation.
What impact could the regulation have on biotech companies and technology markets?
Compliance costs could push consolidation among DNA synthesis vendors, especially smaller companies. Meanwhile, firms focused on sequence analysis, screening systems, and biological threat databases could benefit if federal rules create broader demand for their tools.
How might these regulations affect decentralized science projects involving synthetic biology?
Blockchain-based DeSci projects that rely on open biological data sharing and synthetic biology tooling could face new friction. Mandatory screening at the DNA synthesis layer may force them to add compliance steps to workflows that were designed around open, permissionless access.
