The small African lab sounding the alarm on AI bioweapons
In 2024, researchers at a modest artificial intelligence laboratory in Accra, Ghana, deliberately disabled the safety mechanisms of their drug discovery AI. They then prompted the machine to generate designs for potentially deadly toxins. While they never synthesized these substances themselves, software used to assess their toxicity classified them all as highly dangerous, with some resembling the deadly poison ricin.
The study was published on the arXiv preprint server in May 2025, prompting surprise among the team, including founder Darlington Akogo. Akogo's concern extended beyond the laboratory, questioning the possibility of such designs being obtained by manufacturers of DNA synthesis services. This issue is gaining traction in Africa, with officials like Brenda Obuya from Kenya’s Ministry of Information, Communications and the Digital Economy calling for regulation of DNA synthesis providers.
Akogo has been traveling the globe, warning about the existential threat AI poses in the realm of bioweapons. The study has reached even international bodies, such as Interpol. AI experts, like Microsoft co-founder Bill Gates, have also issued warnings, stating that AI is now capable of empowering bioterrorists to cause mass casualties.
In the wake of the MinoHealth Lab's findings, AI systems have proven capable of outperforming highly trained biologists. The team used a general-purpose language model trained for biology and drug discovery, demonstrating that even general-purpose AI can be manipulated to create hazardous content. This has raised concerns about the increasing accessibility and capabilities of AI.
Written by urgent.news from Rest of World's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.