AI's Existential Threat: How Artificial Intelligence Could Accelerate Bioweapons Development
Technologists and policymakers are increasingly concerned that AI could lower the barrier to creating sophisticated and dangerous pathogens.
The rapid advancement of artificial intelligence is raising alarms among technologists and policymakers, who fear that AI could significantly increase the risk of catastrophic bioweapons and pandemics. While some concerns focus on hypothetical superintelligent AI posing an existential threat, a more immediate worry is AI's potential to accelerate the development and deployment of dangerous biological agents.
Democratizing Bioterrorism
One primary concern is that AI, combined with advances in biotechnology like gene editing and DNA synthesis, could lower the technical barriers to creating dangerous viruses. This could empower a larger number of individuals to assemble known dangerous viruses or even engineer novel ones. Researchers have demonstrated that AI models can provide detailed instructions for creating or modifying pathogens. For instance, a 2023 study showed chatbots offering protocols for engineering treatment-resistant strains of pathogens and advice on biological payload dispersal. More recently, an AI model trained on DNA sequences generated blueprints for deadly viruses, and one company reported thwarting suspected state actor attempts to use its AI for bioweapons research.
These developments are particularly concerning given the increasing accessibility and affordability of gene-editing tools like CRISPR and DNA synthesis. Experiments have shown that it is relatively easy to order fragments of dangerous viral genomes from commercial suppliers, which can then be assembled into functional viruses. AI could further simplify this process by providing step-by-step guidance, troubleshooting, and access to information, potentially enabling individuals with limited lab experience to engage in sophisticated bioengineering.
Engineering Novel and Dangerous Pathogens
Beyond aiding in the replication of known viruses, AI may also help scientists discover or engineer entirely new, more dangerous pathogens. Researchers are exploring the use of AI models, sometimes called genomic language models, to generate novel virus designs. While one study focused on bacteriophages (viruses that infect bacteria), the underlying techniques could theoretically be applied to human viruses. Such AI tools might not only design entirely new viruses but also help modify existing ones to be more transmissible, lethal, or resistant to vaccines and treatments.
This prospect raises fears that advanced AI could assist in creating a "supervirus" more dangerous than anything nature has produced. While the technical hurdles to engineering such a pathogen remain significant, the potential for AI to accelerate discovery and development in this area is a growing concern. Historically, even state-sponsored bioweapons programs, despite their hazards, have led to accidental outbreaks and leaks, underscoring the inherent risks of working with dangerous pathogens.
The Nuance of AI's Role
Despite these concerns, experts note that AI alone does not eliminate the significant practical challenges of biological engineering. Creating dangerous pathogens requires substantial hands-on laboratory expertise, often referred to as "tacit knowledge," which involves skills that are difficult to convey through text alone. Studies suggest that while AI can assist with theoretical or information-gathering aspects of biological challenges, it does not necessarily translate to improved performance in actual lab work for individuals lacking practical experience.
Furthermore, the development of a truly novel and devastating virus is an immensely complex scientific undertaking, with unpredictable outcomes even with AI assistance. The analogy has been drawn between playing a video game simulation of a sport and performing the actual physical activity at an elite level; AI might provide the simulation, but real-world execution remains a formidable barrier.
Preparing for Future Pandemics
Regardless of AI's precise impact on biosecurity, the threat posed by novel pathogens, whether natural or synthetic, remains a significant global security concern. Many of the necessary preparedness measures against AI-accelerated bioweapons overlap with those needed for natural pandemics. Experts advocate for increased government investment in public health infrastructure, including the development of broad-spectrum vaccines and antivirals, enhanced surveillance systems for early outbreak detection, improved personal protective equipment, better ventilation in indoor spaces, and robust screening protocols for synthetic DNA orders.
The conversation around AI's potential risks is intensifying, with some researchers highlighting a greater than 10 percent chance of AI leading to human extinction within a decade. While fears of AI-driven apocalyptic scenarios often capture public attention, addressing the more immediate and tangible threat of AI's role in potentially enhancing bioweapons development and increasing the risk of devastating pandemics is seen as a critical priority.