Researchers from Stanford University and the Ark Institute have successfully used artificial intelligence to design complete genomes for new viruses capable of reproducing and attacking bacteria. This achievement represents a significant advancement, as it moves beyond using AI to design individual proteins or genes to creating entire genetic systems that can function within cells.

Although the word "viruses" may evoke concern, these are not human-infecting viruses; rather, they are bacterial phages, which are specialized viruses that target only bacteria and do not attack human cells. This makes them a promising area of research for finding new solutions to combat antibiotic-resistant bacteria.

The scientific team relied on a specialized AI model called Evo 2, designed to analyze DNA sequences and understand patterns within genes. Unlike AI models that generate text or images, this model works with the "language of life," genetic code, where it can propose new DNA sequences based on the genetic data it has been trained on.

The AI did not directly create the virus but played a role as a genetic designer. It created new genetic blueprints for a well-studied small bacterial phage genome, and scientists then analyzed these designs and selected the most likely sequences to work before testing them in the lab.

The experiment focused on the bacterial phage ΦX174, a small virus that infects bacteria but does not attack human cells. After manufacturing genetic sequences suggested by artificial intelligence and testing them, researchers succeeded in producing 16 effective phages, some of whose copies showed an ability to kill bacteria more efficiently than the original virus from which they were designed.

Scientists believe that this achievement could open the door to developing new treatments based on designing phages tailored to attack specific types of bacteria, especially those that have become resistant to traditional antibiotics. In the future, artificial intelligence may help speed up finding treatments designed according to the nature of each bacterial infection.

However, this scientific development raises questions about biological security at the same time. Experts warn of the need for clear controls and regulations to ensure that these technologies are used in a safe manner as artificial intelligence’s capabilities in dealing with genetic information develop.

Researchers affirmed that the goal of this technology is not to create new risks, but rather to use the power of artificial intelligence to address complex health problems. They said that the future potential of these models is great, provided they are developed responsibly and cautiously.

This achievement is seen as a turning point in the relationship between artificial intelligence and biology, where artificial intelligence has moved from analyzing the secrets of life to contributing to designing parts of it.

The study was published in the journal Science and involved a research team from Stanford University led by Dr. Brian L. Hie, an associate professor of chemical engineering and researcher in the field of biological systems design. The team relied on Evo 2 to generate and analyze DNA sequences and test the feasibility of converting AI-designed genetic designs into effective bacterial phages.

Experiments showed that some genomes designed by artificial intelligence were capable of functioning in a laboratory setting, providing evidence for the potential use of generative models in designing new biological systems. The researchers believe this technology could become an important tool in developing treatments against antibiotic-resistant bacteria in the future, while emphasizing the need to adhere to strict biosafety procedures.

According to the BBC, the significance of this achievement lies not only in producing new phages but also in demonstrating artificial intelligence's ability to move from analyzing genetic code to designing it. The researchers explained that this step represents a shift from using AI to understand existing biological systems in nature to contributing to the invention of new genetic systems that can be tested in a laboratory setting, which could change the future of drug design and biotherapeutic development.

The BBC noted that scientists believe the current development poses a new challenge, as it is no longer just about AI’s ability to understand life, but also its capacity to redesign parts of it. Therefore, experts emphasize that keeping pace with this progress should go hand in hand with developing monitoring and safety systems to prevent the misuse of these technologies.