How the monitored publications were assessed

Discourse valence across every scored item: what the sampled sources published and how independent models read it. It is not a measure of public opinion, not a measure of the state of the world, and article volume is not agreement.

26 items scored in this window

• Anthropic details bad actors’ efforts to misuse its AI for bioweapons
  -84 The Guardian: Artificial Intelligence 2026-09-10 moderate agreement · ±35
• Chinese AI tool told researchers how to make bioweapons
  -76 BBC News: Technology 2026-09-29 low agreement · ±50
• Claude users found ways around safeguards for bioweapons research
  -72 Ars Technica: AI 2026-09-11 moderate agreement · ±45
• ‘These risks are real’: the threat of AI bioweapons - video
  -70 The Guardian: Artificial Intelligence 2026-10-20 low agreement · ±50
• Could AI ‘change the stakes’ in the development of bioweapons?
  -70 The Guardian: Artificial Intelligence 2026-10-14 low agreement · ±55
• Gates says Trump inaction leaves world exposed to AI bioweapons
  -65 Financial Times: Artificial Intelligence 4d ago high agreement · ±20
• ‘These risks are real’: the threat of AI bioweapons – podcast
  -65 The Guardian: Artificial Intelligence 2026-10-20 low agreement · ±55
• Anthropic says it stopped scientists potentially developing bioweapons with AI
  -59 Financial Times: Artificial Intelligence 2026-09-10 moderate agreement · ±40
• Anthropic blocks possible attempt to use AI to make biological weapons
  -58 BBC News: Technology 2026-09-11 low agreement · ±65
• Roundtables: A Conversation With the Creator of AI-Designed Viruses
  -56 MIT Technology Review: AI 2026-10-09 low agreement · ±85
• Google figures out how to watermark AI-designed proteins
  -50 Ars Technica: AI 2026-09-30 single model
• Researchers use AI to design proteins that cut up other proteins
  -42 Ars Technica: AI 2026-10-14 low agreement · ±175
• Lab-grown brain organoids: time for international oversight
  -41 Nature 2026-09-29 low agreement · ±100
• Your AI doomsday questions answered: ‘What, if anything, can people like me do about it?’
  -20 The Guardian: Artificial Intelligence 2026-09-19 single model
• Artificial intelligence and biosecurity: capabilities, threat pathways, and defense-in-depth governance
  -19 arXiv cs.AI 2026-09-16 moderate agreement · ±45
• Single-nucleus transcriptome-wide association study of human brain disorders
  -16 Nature 2026-09-23 low agreement · ±110
• BioEVAL: A global, multi-institutional benchmark of large language and multimodal models for bioengineering
  -5 arXiv cs.AI 2026-09-28 low agreement · ±175
• Do Vision CIL Methods Transfer to Protein? A Diagnostic Benchmark of Frozen Protein Encoders
  0 arXiv cs.LG (Machine Learning) 2026-10-15 single model
• Introducing SynthID Bio
  0 Google DeepMind 2026-09-30 single model
• Introducing Quine: An AI research system designed for the complexity of biology
  +2 Microsoft Research 2026-09-29 low agreement · ±175
• AlphaGenome Atlas: A predictive map of every possible DNA letter change in the human genome
  +7 Google DeepMind 2026-09-08 low agreement · ±175
• How a team of AIs discovered a promising lung-cancer drug
  +12 Nature 2026-09-17 low agreement · ±175
• AI agents run this automated biology lab that sequences DNA and more
  +50 Nature 4d ago single model
• UK to invest £100m in 'world first' AI Biosecurity Lab
  +50 BBC News: Technology 2026-10-24 single model
• Using thousands of AI agents to uncover new leads in biological data
  +50 OpenAI 2026-10-21 single model
• Google DeepMind Maps 9 Billion Possible DNA Variants
  +55 IEEE Spectrum: AI 2026-09-08 moderate agreement · ±40

Across the whole window

The two tables below describe every item in the last 90 days under bio. They are deliberately not narrowed by the reading filter, because a mean taken only over items already selected for their score would just restate the filter.

By topic

By source

A source's mean says how the items it published were read. It is not a rating of the source.