For the first time, scientists say they have used viruses designed by artificial intelligence to treat a human patient.
Researchers announced on Tuesday that they had administered the virus, known as a phage, to a patient suffering from a deadly bacterial infection. Phages attack bacteria, and the hope is that this one will kill the microbes causing the infection.
Radical Numerics, the start-up that created the viruses, posted a technical report about them on its website. But the company declined to provide information about the patient or the doctors running the trial, citing confidentiality requirements.
The report has not been peer reviewed or published in a scientific journal. Company officials said the trial had been authorized by the Food and Drug Administration.
“We take our responsibility to protect patient confidentiality and maintain the trust of our clinical partners extremely seriously,” said Eric Nguyen, the chief executive officer of Radical Numerics.
“We intend to publish our scientific findings through peer-reviewed research, with any future disclosure of clinical details subject to appropriate consent,” he added.
Outside experts greeted the news with curiosity, and a wait-and-see attitude. “If it works, that’s incredible,” said Michael Baym, a microbiologist at Harvard University who was not involved in the research.
If the experiment succeeds, it could point to a new way to use A.I. in medicine, providing individualized treatment that targets even the most resistant bacterial infections.
Dr. Baym said he would reserve judgment until the treatment is finished and it becomes possible to examine just how novel the company’s viruses are. “My impulse is to be skeptical,” he said.
A.I.-designed viruses may be new, but phage therapy dates back over a century.
Now, as antibiotic resistance becomes a growing problem worldwide, the treatment has had a renaissance. In the United States and elsewhere, clinics specializing in phage therapy have opened to help patients struggling against infections that resist multiple drugs.
But this treatment is much harder than popping a few pills of penicillin.
Each type of phage infects only a limited number of bacterial strains. Once doctors get hold of one, they have to screen the virus to see if it has a chance of working against their patient’s infection.
“Hunting for phages is remarkably time-consuming,” said Dr. Marisa Azad, an infectious disease clinician at the Ottawa Hospital in Canada who is working on a collaboration with Radical Numerics.
Some researchers are trying to overcome that barrier by searching in soil and water for new species of phages, enlarging the medical arsenal. But the emergence of A.I. has given scientists another idea: design new phages, customized to fight the particular bacteria infecting a patient.
In 2025, researchers at the Arc Institute in Palo Alto, Calif., created an A.I. model called Evo 2, which they trained on millions of DNA sequences, including phage genes. They prompted Evo 2 to make new genomes similar to those in a group of phage species.
The A.I. generated 700,000 virus genomes. The team synthesized DNA molecules from 285 genomes and inserted them into bacteria. The researchers found that 16 of the genomes gave rise to viable viruses.
Three members of the Evo team, including Dr. Nguyen, left to found Radical Numerics. In June, the company unveiled a new A.I. model, called Omnii, based on some of the same principles as Evo.
Omnii can be trained not just on DNA but also on other aspects of biology, such as the structures of proteins and molecules that keep genes turned on or off.
The company is now using Omnii to tackle a range of challenges, including designing cancer vaccines and recognizing the risk that a virus may cause a pandemic. Among these efforts is the first trial of A.I.-designed phage therapy.
The researchers created variations of a phage called T7. The virus, which only infects E. coli, looks like a microscopic Apollo lunar lander: a ball-shaped protein shell that holds its DNA atop six leglike protrusions.
To infect a bacterium, T7 lands on its surface, using its legs to anchor in place. It then injects its DNA into the cell.
Omnii generated new viral genes altering the shapes of T7’s legs, according to Dr. Nguyen and his colleagues. While some of the new legs were significantly different from natural ones, Omnii’s phages could still infect bacteria.
Once the researchers were confident their strategy had worked, they partnered with doctors treating a patient with a stubborn E. coli infection. The bacteria had resisted not just antibiotics but all the available phages.
The researchers said they had trained Omnii with information about the resistant strain and told it to design T7 phages with legs that could grab onto the bacteria.
Once the A.I. generated genomes for new viruses, the scientists said they produced the corresponding DNA molecules and inserted them in bacteria. The bacteria became infected and produced T7 phages that could successfully infect, and kill, other microbes.
After running tests to make sure the phages were safe, Radical Numerics presented them to the doctors. The F.D.A. authorized a trial.
Brandon Berryhill, an evolutionary microbiologist at Uppsala University in Sweden, said there was no guarantee that the A.I.-designed phages would work. “We cannot measure anything in the lab and say how good it’s going to be at controlling something in an infection,” he said.
Even if the treatment does work, simply creating phages with the help of A.I. won’t help scientists to understand what determines why a phage therapy succeeds or fails. “There’s just so much we don’t know about the rules of the game,” Dr. Berryhill said.
It will take a while to know if Omnii’s novel phages succeed. Phage therapy can last for days or weeks, and doctors typically monitor their patients for months in case the infection rebounds.
If the trial is successful, Dr. Nguyen hopes the company can help speed phage therapy by cutting down the time it takes to create a new virus tailored to each patient’s infection.
“We think we can get it done in days,” he said.
The post Start-Up Claims to Treat a Patient With A.I.-Created Viruses appeared first on New York Times.
