Subvolts take AI accelerates dangerous biological research; new frameworks aim to control risky data. Mar 11, 2026 - Developers
- Understand how AI tools can be misused in sensitive research areas and the ethical considerations involved.
- Business
- Learn about emerging risks in AI-driven biological research and potential policy responses.
- Everyday users
- Be aware of how AI advancements could impact public health and biosecurity risks.
This is a solid interview with a researcher discussing critical AI safety concerns in a sensitive field.
Bioinfohazards: Jassi Pannu on Controlling Dangerous Data from which AI Models Learn
Audio streams from Erik Torenberg, Nathan Labenz. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Jassi Pannu discusses how AI advancements in biological research increase the risk of engineered pandemics. The episode covers biosecurity, AI's ability to bypass data safeguards, and a proposed framework to restrict dangerous biological data while supporting open science. It also explores defense strategies like DNA synthesis screening and pathogen surveillance.
AI tools are speeding up biology research, but this also makes it easier to create dangerous viruses. The discussion explores ways to control the riskiest biological information while keeping science open, and how to protect against future threats.
This episode is for researchers, policymakers, and anyone concerned about AI's dual-use potential in biology. It connects directly to frontier labs' work by highlighting how powerful AI models can be misused, even unintentionally, by accelerating dangerous research. Pannu's proposed data framework and defense strategies offer concrete ideas for mitigating these risks, relevant to ongoing discussions about AI safety and regulation.
- HostNathan Labenz
- GuestJassi PannuAssistant Professor, Johns Hopkins
- Guest: Jassi Pannu
- Title: Bioinfohazards: Jassi Pannu on Controlling Dangerous Data from which AI Models Learn
- Show: The Cognitive Revolution
- Length: 103 minutes
- Topic: AI transforming biological research and raising risk of engineered pandemics
- Proposed Framework: Biosecurity Data Level framework
Jassi Pannu, Assistant Professor at Johns Hopkins, explains how rapidly advancing AI is transforming biological research and raising the risk of engineered pandemics. They map today’s biosecurity landscape, from pathogen detection and DNA sequencing to vaccine development, and examine how frontier models can already troubleshoot lab work and bypass data safeguards. The conversation introduces a proposed Biosecurity Data Level framework to restrict only the most dangerous functional biological data while preserving open science. They close with a broader defense-in-depth strategy—Delay, Deter, Detect, Defend—including DNA synthesis screening, global pathogen surveillance, and practical tools like Far UV sterilization.
LINKS:
Jassi Pannu: https://x.com/JassiPannuMD
Science article that prompted this conversation: https://www.science.org/doi/10.1126/science.aeb2689
Sponsors:
VCX:
VCX, by Fundrise, is the public ticker for private tech, giving everyday investors access to high-growth private companies in AI, space, defense tech, and more. Learn how to invest at https://getvcx.com
Framer:
Framer is an enterprise-grade website builder that lets business teams design, launch, and optimize their.com with AI-powered wireframing, real-time collaboration, and built-in analytics. Start building for free and get 30% off a Framer Pro annual plan at https://framer.com/cognitive
Claude:
Claude is the AI collaborator that understands your entire workflow, from drafting and research to coding and complex problem-solving. Start tackling bigger problems with Claude and unlock Claude Pro’s full capabilities at https://claude.ai/tcr
Tasklet:
Tasklet is an AI agent that automates your work 24/7; just describe what you want in plain English and it gets the job done. Try it for free and use code COGREV for 50% off your first month at https://tasklet.ai
CHAPTERS:
(00:00) About the Episode
(05:59) From outbreak to vaccine
(17:08) Threat actors and data (Part 1)
(21:23) Sponsors: VCX | Framer
(23:53) Threat actors and data (Part 2)
(31:05) Gain-of-function research risks (Part 1)
(37:39) Sponsors: Claude | Tasklet
(41:03) Gain-of-function research risks (Part 2)
(48:05) AI models in biology
(01:00:51) Dangerous AI capabilities
(01:07:59) Biosecurity data level framework
(01:18:58) Policy, governance, and infrastructure
(01:28:53) Defense in depth vision
(01:40:43) Episode Outro
(01:45:02) Outro
PRODUCED BY:
- Bioinfohazards
- Jassi Pannu
- Risks arising from the intersection of biology and information technology, especially AI.
- Engineered pandemics
- Jassi Pannu
- Disease outbreaks intentionally created or modified using technology.
- Biosecurity
- Jassi Pannu
- Measures taken to protect against the accidental or intentional release of biological agents.
- DNA synthesis
- Jassi Pannu
- The process of artificially creating DNA sequences.
1 h 27AI:AM: Where's the Steering Wheel? My Agent Set the Password, Hidden Valence, Proofs Find Linux Bugs
WithCameron Berg, Everett Hildenbrandt, Grigore Rosu, Phil Maher
2 h 192 h 021 h 28AI:AM: A Level We Shouldn't Pass? Notes from The Curve + Tokens vs. Salaries & Is SaaS Cooked?
WithThomas Sohmers, swyx, Evan Miyazono, Edward Hu
Questions people ask
- What is the main concern discussed regarding AI and biology?
- The rapid advancement of AI in biological research increases the risk of engineered pandemics by potentially enabling easier creation of dangerous pathogens.
- What is the proposed solution to manage dangerous biological data?
- A Biosecurity Data Level framework is proposed to restrict only the most dangerous functional biological data while preserving open science principles.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Erik Torenberg, Nathan Labenz.