AI Can Be Used To Develop Biological And Chemical Weapons

Authored by David Gangemi via RealClearDefense,

The greatest threat to America may be something we cannot see.

Biological and chemical weapons can spread debilitating fear and wreak havoc on populations. Unfortunately, the threat of these kinds of attacks on American soil is only rising, empowered by the unique capabilities of artificial intelligence. To prevent catastrophe, we must use AI to counter the danger that this advanced technology can unleash.

We are only one misstep away from catastrophe. The most recent close call came last month when two Chinese nationals, who received funding from their communist government to work on a noxious fungus, were caught smuggling the pathogen into the United States. The FBI arrested them just in time, as this was no ordinary fungus—instead, it was what scientists call an “agroterrorism weapon” that would have decimated America’s grain farms, sickened the U.S. population, and disrupted our nation’s food supply.

Those who lived through the fraught days immediately after 9/11 likewise remember the anthrax scare, as toxic letters were sent through the postal service, killing five people and making everyday Americans terrified to open their mailboxes. 

Every few years, some new suspect threatens our military bases, political leaders, or someone else with ricin, a deadly poison derived from the castor plant.

And just a few short years ago, millions died and the entire world was thrown into a tailspin when COVID-19—which many experts now believe originated from questionable handling and a lab leak at the Wuhan Institute of Virology—crossed borders with abandon.

The rapid rise of AI is only making this problem more prevalent. In 2022—months before ChatGPT was released, bringing large language models to the masses—an AI designed to develop drugs invented 40,000 new chemical weapons in a mere six hours. In 2023, AI was used to provide a recipe for poisonous chloramine gas, which it called “Aromatic Water Mix.” AI experts and government officials have been warning for years that AI could spark the next pandemic and be an unparalleled tool in the hands of terrorists.

These facts are encouraging some to ask that AI developers give pause to the negative consequences of this powerful technology. We have enough problems with biological espionage, terrorism-by-mail, and lethal lab leaks. Why would we put potential biological and chemical weapons generators in the hands of anyone with a computer?

But responding to this threat is not so simple as pulling the plug. First, while AI has the potential to be used for evil, it also has immense power for good. The same tools that could be used to make biological weapons are also being applied to cure currently untreatable diseases. Additionally, America can’t stop others from developing AI for whatever uses they desire. COVID-19 and the recent agroterrorism fungus both came from China, and you can bet China will have no problem unleashing AI for even more destructive ends if it serves their interests, as will every other bad actor in the world.

So what else can we do?

First, the administration should continue to aggressively investigate and thwart potential acts of biological and chemical terrorism. The recent FBI arrest of the Chinese fungus smugglers proves that America’s law enforcement is aware of this threat and still capable of preventing attacks before they happen.

Likewise, President Donald Trump acted presciently in his first administration by launching the first-ever National Biodefense Strategy in 2018, which outlined how our nation can defend against natural outbreaks and intentional biological attacks. This strategy, coupled with the president’s swift military action against Syria for its use of biological weapons that same year, reveals that the current administration will still use the immense power of American deterrence to stop the use of these deadly weapons. 

Yet with AI poised to rapidly exacerbate biological and chemical weapon proliferation, traditional tools are not enough. We must use AI to respond to AI.

The private sector is already on the case. Companies like Renovaro, and OpenAI, two U.S.-based AI firms, are already applying machine learning to both prevent AI from producing recipes for weaponry and to identify and counter biological and chemical threats before they can spiral out of control. 

New coding produced by top AI experts can be applied to any large language model to prevent it from teaching users how to make weapons of mass destruction. For those pathogens that slip through the cracks, Renovaro’s AI has the potential to develop antidotes to biological and chemical samples within five days, lightyears faster than the long months it took to develop vaccines and effective treatments for COVID-19.

President Trump promised a golden dome to protect America from missile attacks—a worthwhile initiative. Yet the next war may be started not by a missile, but by a microbe. We need a biological golden dome, and AI can make it.

Dr. David Gangemi is a Professor Emeritus at Clemson University, Former Senior Science Advisor on health affairs to the Assistant Secretary of Defense, and a virologist at the U.S. Army Medical Research Institute of Infectious Diseases