Artificial intelligence can now autonomously design and run thousands of biological experiments without human intervention, but the regulatory systems meant to oversee these capabilities are falling dangerously behind. In February 2026, OpenAI and Ginkgo Bioworks announced that OpenAI's GPT-5 model had independently designed and executed 36,000 biological experiments through a robotic cloud laboratory, reducing the cost of producing a desired protein by 40 percent . The breakthrough highlights a critical vulnerability: current safety measures and international treaties were never designed to account for AI-driven automation in biology.

How Does AI Actually Run Lab Experiments on Its Own?

The process works like a tight feedback loop between artificial intelligence and robotics. A human sets a biological goal, such as designing a new protein or optimizing a cellular process. The AI model then proposes experimental designs, and robots in a cloud laboratory execute those designs automatically. The results feed back to the AI, which learns from the data and proposes the next round of experiments. This cycle repeats thousands of times in weeks, compressing what would traditionally take months or years of human trial-and-error into days .

This represents a fundamental shift in how biology works. For decades, the field moved from observation toward understanding, with scientists sequencing genomes to catalog DNA and learning how genes encode proteins. Tools like CRISPR then allowed targeted DNA editing. Now, AI closes the loop by both designing biological systems and rapidly testing them at scale .

What Are the Biggest Safety Risks Nobody's Prepared For?

The speed and autonomy of AI-driven biology create what researchers call the "dual-use problem": technologies developed for beneficial purposes can be repurposed to cause harm. Stephen D. Turner, a data scientist at the University of Virginia who studies genomics and biosecurity, explained the core challenge .

"Current safety measures and regulations have not kept pace with these capabilities, and the gap between what AI can do in biology and what governance systems are prepared to handle is growing," Turner stated.

Stephen D. Turner, Data Scientist, University of Virginia

The risks span multiple domains. AI models integrated with automated labs can optimize how well a virus spreads, even without specialized training. Researchers have determined that AI could lower barriers at multiple stages in developing bioweapons, such as altering which species a pathogen infects or helping it evade the immune system . Current AI models can even walk users through the technical steps of recovering live viruses from synthetic DNA.

Perhaps most concerning, studies show that AI tools can dramatically reduce the expertise barrier to dangerous work. When researchers from Scale AI and SecureBio gave people with limited biology experience access to large language models (LLMs), which are AI systems trained on vast amounts of text data, those novices completed biosecurity-related tasks with four times greater accuracy than their baseline performance. Around 90 percent of these inexperienced users reported little difficulty getting the models to provide risky biological information, such as detailed instructions on working with dangerous pathogens, despite built-in safety filters .

Why Are Existing Regulations Completely Unprepared?

The regulatory landscape reveals a troubling mismatch. Rules governing biological research do not account for AI-driven automation, and rules governing AI do not specifically address its use in biology . The 1975 Biological Weapons Convention, an international treaty prohibiting the production and use of bioweapons, contains no provisions for AI whatsoever.

In the United States, the Biden administration issued a 2023 executive order on AI security that included biosecurity provisions, but the Trump administration revoked it. Screening the synthetic DNA that commercial providers manufacture to ensure it cannot be misused remains mostly voluntary. A bipartisan bill introduced in 2026 to mandate DNA screening does not yet address AI-designed sequences that evade current detection methods .

The safety evaluations that AI labs run before releasing new models are often opaque and unsuited to capture real-world risk. Researchers have estimated that even modest improvements in an AI model's ability to help plan pathogen-related experiments could translate to thousands of additional deaths from bioterrorism per year .

What Steps Are Experts Proposing to Close the Governance Gap?

• Managed Access Frameworks: The Nuclear Threat Initiative has proposed matching who can use a given AI tool to the risk level of the model rather than imposing blanket restrictions, allowing beneficial research while limiting dangerous access.
• Improved DNA Synthesis Screening: The RAND Center on AI, Security and Technology outlined actions including enhanced screening of DNA sequences before synthesis and model evaluations before AI systems are released to the public.
• Biological Data Governance: Researchers have argued that biological data itself needs governance, especially genomic data that could train AI models with dangerous capabilities.
• Coordinated Government Action: The UK AI Security Institute and the US National Security Commission on Emerging Biotechnology have both called for coordinated international government action to establish clear oversight mechanisms.

Some AI companies have begun voluntarily imposing their own safety measures, though critics argue this patchwork approach is insufficient given the stakes .

What's Happening in the Commercial Biology Space?

While governance struggles to catch up, the commercial sector is racing ahead with integrated "sample to insight" solutions. QIAGEN, a global leader in molecular diagnostics and life sciences tools, announced at the American Association for Cancer Research Annual Meeting 2026 that it will showcase new oncology workflow applications combining sample preparation, multi-omics profiling, and AI-powered data interpretation .

The company is introducing the QIAGEN Discovery Platform, described as an "AI-grounding solution" designed to bring together biological knowledge, omics data, and advanced analytics to support oncology research. This platform represents the next generation of tools that pair AI with laboratory automation, enabling researchers to move from raw biological samples to actionable insights with minimal human intervention .

"Cancer research and molecular diagnostics are increasingly constrained by fragmented workflows, variability in sample processing and the growing complexity of multi-omics data," explained Nitin Sood, Senior Vice President and Head of Product Portfolio and Innovation at QIAGEN.

Nitin Sood, Senior Vice President and Head of Product Portfolio and Innovation, QIAGEN

QIAGEN is also rolling out the QIAsymphony Connect, an automated platform for clinical molecular testing that standardizes nucleic acid extraction and improves laboratory productivity. The company has already placed over 3,300 of its established QIAsymphony systems in laboratories worldwide, and the new Connect version promises to scale these capabilities further .

How Should Scientists and Policymakers Respond?

The core tension is clear: AI-driven biology offers enormous benefits for drug discovery, vaccine development, and disease prevention. The same technology that can accelerate protein engineering to respond to emerging infections can also be weaponized. The question is not whether to pursue these capabilities, but how to govern them responsibly.

Turner emphasized that the timeline for action is narrowing. As cloud laboratories become cheaper and more accessible, researchers will increasingly send AI-generated experimental designs to remote facilities for execution, further decoupling the design phase from human oversight . Without coordinated governance frameworks, the gap between what AI can do in biology and what oversight systems can monitor will only widen.

The challenge ahead requires collaboration across governments, AI companies, biological researchers, and security experts to establish clear rules for AI-designed sequences, mandatory DNA screening standards, and transparent safety evaluations before new AI models are deployed in biological research. The 36,000 experiments GPT-5 ran in weeks represent both the promise and the peril of autonomous AI in biology. How quickly regulators can catch up may determine whether this technology becomes a tool for healing or harm.