English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

As advancements in artificial intelligence (AI) propel progress in the lifesciences, they may also enable the weaponisation and misuse of biologicalagents. This article differentiates two classes of AI tools that could posesuch biosecurity risks: large language models (LLMs) and biological designtools (BDTs). LLMs, such as GPT-4 and its successors, might provide dual-useinformation and thus remove some barriers encountered by historical biologicalweapons efforts. As LLMs are turned into multi-modal lab assistants andautonomous science tools, this will increase their ability to supportnon-experts in performing laboratory work. Thus, LLMs may in particular lowerbarriers to biological misuse. In contrast, BDTs will expand the capabilitiesof sophisticated actors. Concretely, BDTs may enable the creation of pandemicpathogens substantially worse than anything seen to date and could enable formsof more predictable and targeted biological weapons. In combination, theconvergence of LLMs and BDTs could raise the ceiling of harm from biologicalagents and could make them broadly accessible. A range of interventions wouldhelp to manage risks. Independent pre-release evaluations could help understandthe capabilities of models and the effectiveness of safeguards. Options fordifferentiated access to such tools should be carefully weighed with thebenefits of openly releasing systems. Lastly, essential for mitigating riskswill be universal and enhanced screening of gene synthesis products.

Artificial intelligence and biological misuse: Differentiating risks of language models and biological design tools