Scientists have demonstrated a major breakthrough in biotechnology by using artificial intelligence to generate an entire microbial genome. A powerful AI model called Evo2 was trained on an enormous dataset of about 9 trillion DNA letters, allowing it to learn patterns in genetic sequences in the same way language models learn patterns in text. When researchers gave the system a portion of a microbe’s genome, the AI successfully generated a plausible version of the remaining genetic code.
The experiment showed that AI can now “write” genetic instructions rather than simply analyze them. In one demonstration, the system generated DNA sequences that even encoded hidden Morse code messages inside chromosome-like structures, illustrating how precisely the model could manipulate genetic information. This ability suggests that AI could one day design biological systems in a highly controlled and creative way.
However, the researchers did not attempt to create a living organism from the AI-generated genome. DNA is extremely sensitive to errors, and even small mistakes can prevent a microbe from surviving. Scientists therefore emphasize that although the sequences looked realistic, they may not produce a viable organism yet. Much more testing and refinement will be needed before AI-generated genomes could function reliably in living cells.
Even with these limitations, experts see the work as a milestone in synthetic genomics, comparing it to a “ChatGPT moment” for biology. If the technology continues to improve, AI could help scientists design microbes that produce medicines, break down pollution, or fight antibiotic-resistant bacteria. At the same time, the development raises important ethical and biosafety questions about how such powerful tools should be regulated.
