Urgent.News

What's breaking now, across thousands of outlets.

AI

Human receives AI-designed virus for the first time

Not to kill the human, of course – to kill the antibiotic-resistant bacteria causing their infection

Human receives AI-designed virus for the first time

For the first time, a human has received a virus designed by artificial intelligence, specifically targeting bacteria. Radical Numerics, an AI lab based in San Francisco, made this claim in a recent announcement. The treatment involved a cocktail of bacteriophages, viruses that infect bacteria, which were created using an AI model called Omnii-1.

Phages have gained attention as potential alternatives to antibiotics due to their ability to target specific bacterial strains, sparing beneficial gut bacteria that can be affected by broad-spectrum antibiotics. The company claims that the AI-designed phages eliminated the patient's antibiotic-resistant bacterial isolate in a lab setting.

However, they have not disclosed the results of administering the treatment to the patient under a compassionate-use arrangement, citing privacy concerns. The company also noted that a single patient's experience does not establish the safety or efficacy of the AI phage therapy. The research write-up does mention that Omnii-1 was able to predict which phages could infect Pseudomonas aeruginosa, a bacterium known for developing antibiotic resistance and causing infections in patients with burns, wounds, and persistent lung infections due to cystic fibrosis.

Written by urgent.news from The Register's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Also reported by 1 other outlet

Read the original at theregister.com →

More in AI

SlowMate AI: Do Less. Move Forward.

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady. What I Built Meet SlowMate AI , a local-first productivity assistant designed for people who feel…

  • SlowMate AI simplifies productivity by generating action plans from user input
  • Uses locally hosted Qwen2.5 3B model for privacy and control
  • Deployed on GitHub Pages with HTML/CSS/JS frontend

More from Thursday 22 October →