Can ChatGPT Create New Coronavirus? OpenAI Worries AI Will Invent Viruses!

Recently, OpenAI's statement exploded like a thunderclap in the tech world and public discourse—they warned that future AI models, such as ChatGPT's successors, might actually enable ordinary people to "create" viruses! This news sent chills down everyone's spine. Are we really about to face AI's "darkest hour"? Is this an overreaction from tech giants, or the harsh reality of artificial intelligence's double-edged sword that we must confront?

AI's "Novice Elevation": When Ordinary People Can Master "Forbidden Arts"

OpenAI's warning is not baseless. They introduced a core concept called "Elevation of Novices." Simply put, artificial intelligence has the ability to instantly arm those "amateurs" who lack professional scientific backgrounds into "experts" capable of manipulating sensitive biological data. Imagine if someone who knows nothing about biology could obtain detailed steps for replicating or modifying viruses through AI's precise guidance, including experimental methods and required materials—this would undoubtedly open Pandora's box.

Although OpenAI's safety systems head Johannes Heidecke stated that this threat hasn't fully materialized yet, future models like successors to O3 and O4-mini are highly likely to provide such capabilities. Moreover, despite strict regulation of laboratories and biological materials, they are not "impregnable fortresses." This "AI-empowered" low-threshold operation has dramatically escalated risks in the biosecurity field, no longer the "privilege" of a few top scientists, but a potential threat that could spread to any corner.

OpenAI's "Self-Rescue" and AI's Dual Nature

Facing such severe challenges, as ChatGPT's "creator," OpenAI naturally doesn't dare to be complacent. They have been preparing for rainy days, continuously strengthening safety measures since April 2025. Now, ChatGPT models can actively intercept requests involving virus creation or biological weapons. Once the system detects suspicious instructions, it immediately triggers algorithmic interception and hands them over to human teams for risk assessment. Additionally, before deploying any new AI model, OpenAI conducts strict testing and collaboration with external experts and public institutions, striving to minimize risks.

However, AI doesn't only have a dark side of "virus creation." OpenAI also emphasizes that the other side of this double-edged sword is the light of hope illuminating human health. These AI models that may pose risks also contain enormous potential to drive medical revolutions. They can accelerate drug discovery, design more efficient vaccines, and even bring hope for curing rare diseases. Imagine AI working day and night in laboratories, contributing its unparalleled computing and analytical capabilities to human health—what an inspiring scene!

Building Defenses Together: Shared Responsibility of Tech Giants, Scientists, and the Public

To better harness this powerful force, OpenAI is actively collaborating with biosecurity teams, Los Alamos National Laboratory, and other top institutions to jointly test AI systems' robustness and study AI's impact in real laboratory environments. The upcoming Biodefense Summit in July 2025 will also bring together researchers, NGOs, and government agencies worldwide to discuss AI's risks and opportunities in the biological field. This is not just OpenAI's responsibility alone, but a shared challenge for all humanity.

Harnessing AI: Where Do We Go From Here?

ChatGPT's "virus creation" alarm undoubtedly sounds a wake-up call for everyone immersed in AI miracles. It forces us to face the reality that technological progress always comes with uncertainty and potential risks. The future of artificial intelligence doesn't lie in whether we can completely eliminate its negative impacts, but in whether we can build strong enough "guardrails" to enjoy the convenience and progress it brings while effectively avoiding potential harm.