Same Question, Three AI Models, Different Answers: A Case Study in AI Biosecurity Safeguards by Nnaemeka Emmanuel Nnadi

Biomed­i­cal re­search is now open­ing up new pos­si­bil­ities thanks to re­cent ad­vances in ar­tifi­cial in­tel­li­gence. Scien­tists are now able to use large lan­guage mod­els (LLMs) to an­swer com­plex sci­en­tific ques­tions, such as vac­cine de­sign. At the same time, con­cerns about pos­si­ble mi­suse of AI in biol­ogy have prompted de­vel­op­ers of cut­ting-edge AI sys­tems to put pro­tec­tions in place to pre­clude mod­els from helping with tasks that could en­able haz­ardous biol­ogy ...