Open-weight genome AI: why data filtering isn't enough

Kari Jaaskelainen

25 Nov 2025 — 1 min read

Can we keep open-weight genome AI models safe just by removing pathogen sequences from their training data? A new study tests that idea and finds it's not enough.

Researchers assessed a state-of-the-art genomic language model and showed that post-release fine-tuning on sensitive pathogen data can partially restore risky capabilities.
The fine-tuned model performed better on unseen viral sequences and even generalized to spot immune-evading variants, despite not seeing those exact viruses during fine-tuning.
Bottom line: dataset filtering alone is brittle for safeguarding open-weight biology models.

The authors urge layered safeguards: stronger model evaluations, clearer release norms for open weights, and technical mitigations that hold up under adversarial fine-tuning.

Paper: https://arxiv.org/abs/2511.19299v1

Paper: https://arxiv.org/abs/2511.19299v1

Register: https://www.AiFeta.com

#Genomics #AI #Biosecurity #ResponsibleAI #OpenSource #ML #Safety

Read more

Meet GRACE: a moral governor for safer, more transparent AI

Meet GRACE: a moral governor for safer, more transparent AI

AI agents are getting powerful—so how do we make sure they do the right thing, not just the effective thing? Meet GRACE, a reason-based moral governor that keeps AI behavior aligned with human norms by separating moral reasoning from goal-driven decision-making. * Moral Module: uses deontic logic and explicit reasons

Why some fine-tuned LLMs miss phishing—and how to fix it

Why some fine-tuned LLMs miss phishing—and how to fix it

Not all fine-tuned LLMs spot phishing equally. A new study tests Llama 3.1 8B, Gemma 2 9B, and Mistral on high-stakes phishing detection—and uses SHAP and mechanistic interpretability to reveal why models do (or don’t) generalize. * Architecture × data diversity matters: Gemma 2 9B hits state-of-the-art performance (F1

AI that explains itself—by following the science

AI that explains itself—by following the science

AI that explains itself—by following the science Black-box AI can be powerful, but it often can’t tell us why it made a decision. Concept Bottleneck Models (CBMs) try to fix that by predicting human-understandable concepts first, then the final answer. The catch: standard CBMs ignore domain-specific cause-and-effect and

Meet GenomAgent: A Team of AI Specialists for Smarter Genomics Q&A

Meet GenomAgent: A Team of AI Specialists for Smarter Genomics Q&A

TL;DR: Finding reliable answers in genomics is hard. GenomAgent turns one big AI into a coordinating team of specialists—and beats the current leader by 12% on a key benchmark. Genomic facts live across many databases. Standard chatbots struggle because they can’t flexibly query those sources. GeneGPT added