研究人員利用人工智慧設計功能性病毒基因組
Researchers Use AI to Design Functional Viral Genomes
Updated at: August 11, 2026 at 12:30 AM
在一項於 2026 年 8 月發表於《科學》雜誌的里程碑研究中，來自史丹佛大學與弧研究院(Arc Institute)的研究人員在合成生物學領域取得了突破性進展。
In a landmark study published in Science in August 2026, researchers from Stanford University and the Arc Institute have achieved a breakthrough in synthetic biology.
在 Brian Hie 的帶領下，該團隊利用了一種名為 Evo 的專門 AI 模型，從零開始設計出功能齊全的病毒基因組。
Led by Brian Hie, the team utilized a specialized AI model called Evo to design fully functional viral genomes from scratch.
在篩選數千個 AI 生成的設計後，研究人員合成了 285 個候選對象，最終產生了 16 種能夠感染並消除大腸桿菌的病毒。
After screening thousands of AI-generated designs, the researchers synthesized 285 candidates, resulting in 16 viruses capable of infecting and eliminating E. coli bacteria.
其中一些合成病毒的功效甚至超越了天然對應物。
Some of these synthetic viruses even surpassed natural counterparts in efficacy.
雖然這項進展為噬菌體療法提供了巨大的潛力——即提供了一種對抗抗生素抗藥性「超級細菌」的方法——但它也帶來了顯著的生物安全疑慮。
While this advancement holds immense potential for phage therapy—offering a way to combat antibiotic-resistant superbugs—it also presents significant biosecurity concerns.
研究人員強調，該 AI 並非自主運行；人類的監督在過程中仍然至關重要。
The researchers emphasized that the AI is not autonomous; human oversight remains central to the process.
隨著生成式生物學的成熟，這項研究凸顯了一個兩用困境：同樣能透過醫學創新拯救無數生命的技術，也需要強大且全新的安全與監管框架，以防止潛在的濫用。
As generative biology matures, this research underscores a dual-use dilemma: the same technology that could save countless lives via medical innovation also demands robust new safety and regulatory frameworks to prevent potential misuse.
