An AI Invented 40,000 Potential Chemical Weapons in Just Six Hours

A team of researchers caused worry as they used Artificial Intelligence –AI– to invent 40,000 potential lethal weapons that took approximately six hours, The Verge reported.

AI facilitates daily living, paving the way for self-driving cars, increasing home security, assisting in preventing fraud and so. The technology in question developing at an unprecedented pace, however, has negative sides.

The results of the newly conducted research indicated how AI used for drug study could be misemployed for the sake of producing chemical weapons. The research has been published in Nature Machine Intelligence.

What to know about VX

As part of the study, the scientists used some drug-like molecules to train the AI. “To narrow the universe of molecules, we chose to drive the generative model towards compounds such as the nerve agent VX,” they said. VX is known to be a tasteless and odorless nerve agent — one of the most toxic chemical warfare agents developed in the UK.

“In less than 6 hours after starting on our in-house server, our model generated 40,000 molecules that scored within our desired threshold,” said the scientists. “In the process, the AI designed not only VX but also many other known chemical warfare agents that we identified through visual confirmation with structures in public chemistry databases”.

Credit: Nature Machine Intelligence

As such possible lethal weapons could pose a catastrophic risk to human beings, the scientists involved in the study were warned not to share some critical details of it.

“The way VX is lethal is it actually stops your diaphragm, your lung muscles, from being able to move so your lungs become paralyzed,” Fabio Urbina, lead author of the paper, told The Verge.

“The biggest thing that jumped out at first was that a lot of the generated compounds were predicted to be actually more toxic than VX,” said Urbina. “And the reason that’s surprising is that VX is basically one of the most potent compounds known”.

The scientists are concerned

In their study, the researchers underscored that the outputs concerning biological agents should be a kind of wake-up call for their colleagues who resort to Artificial Intelligence when discovering drugs.

Also, Urbina uttered his concerns as he said that the dataset was open to everyone for free. He further stated, “You can go and download a toxicity dataset from anywhere. If you have somebody who has some machine learning capabilities, they could build something like this generative model driven by toxic datasets”.