A growing number of technologists and policymakers are expressing concerns that artificial intelligence (AI) could pose significant risks to humanity, including the potential to contribute to the creation of synthetic pandemics. Last week, a leading researcher at Anthropic, an AI company, suggested there is a greater than 10% chance that AI could cause the extinction of all humans within the next decade. This warning has led to calls for slowing down the development of AI, and both Anthropic and OpenAI have acknowledged these concerns. One major worry is that AI could make bioweapons more powerful and easier to create by accelerating advances in biotechnology. This doesn’t necessarily require AI to become superintelligent, but rather to enhance existing tools like gene editing and DNA synthesis. Chatbots, for example, could help more people assemble dangerous viruses that are already known to science, increasing the risk that someone with malicious intent could develop a catastrophic bioweapon. Recent events have added weight to these concerns. Earlier this year, biologists warned that chatbots could provide detailed instructions on how to engineer a pathogen resistant to treatment or even recreate a virus that caused a past pandemic. In August, scientists discovered that an AI model trained on DNA sequences had generated blueprints for viruses not found in nature, although these were only lethal to bacteria. Anthropic also reported that it had blocked attempts by state actors to use its AI for bioweapons research. The risk of an artificial pandemic is still unclear, but advances in synthetic biology have made bioengineering tools more powerful and accessible. Even with AI, there are still significant technical challenges in creating novel, highly dangerous viruses. Synthetic superviruses are particularly concerning because infectious diseases have historically been among the deadliest forces on Earth. A bioterrorist could, in theory, engineer a virus that is both highly contagious and deadly, surpassing the natural constraints of evolution. Such a virus could have catastrophic effects on society, potentially leading to widespread disruption and collapse. Experts like MIT biologist Kevin Esvelt warn that breakthroughs in both AI and synthetic biology are expanding the number of people who could engineer known viruses. The cost and difficulty of genome editing have dropped significantly, and DNA synthesis has become more affordable. However, dangerous biological materials are often easily accessible, as shown by an experiment where a student ordered fragments of the 1918 influenza genome using a pseudonym, and most companies sent the pieces without verification. AI may further reduce the barriers to viral engineering by providing step-by-step instructions and troubleshooting help. In 2023, MIT researchers found that chatbots could guide users in engineering a pandemic, suggesting pathogens, protocols, and even research organizations. More recent studies show that AI models have outperformed expert virologists in solving complex lab problems. Despite these concerns, the risk of a synthetic pandemic remains uncertain. While AI may assist in some aspects of bioweapon development, the practical challenges of engineering viruses remain significant. Even with advancements in CRISPR and AI, biology remains a complex and unpredictable field. To address these risks, governments are urged to invest in pandemic preparedness, including vaccines that protect against multiple virus families, broad-spectrum antivirals, and improved surveillance systems. Enhancing indoor environments to reduce the spread of airborne viruses and implementing strict screening for synthetic-DNA orders are also recommended. While fears of superintelligent machines dominate public discourse, attention should also be given to the real and existing threats posed by synthetic pandemics.