Experts warn of threats of AI bioterrorism

Semafor · collected 2026-09-06 · by Tom Chivers
Read the original at Semafor ↗

Summary

Experts in the field of artificial intelligence (AI) are warning about the potential threats of AI bioterrorism, citing the dual-use nature of powerful AI models in bioscience that can be used to create both vaccines and bioweapons. A Google DeepMind researcher noted that information used to develop a vaccine could also be misused to create a bioweapon. According to an immunologist quoted in the article, it is only a matter of time before an accident occurs due to AI misuse in bioscience. The threat of AI-driven bioterrorism is seen as particularly concerning given that terrorist groups are already using AI technology.
Written by the local model on 2026-09-06, using this article's own text rather than the other coverage of the same event (that is the story summary below).

Signals How these are calculated →

Claims extracted
6
claim-shaped sentences
Uncertain
0%
0 of 6 hedged
Leaning
Leans right
of the writing, not the subject
Publisher trust
96.2
red-flag proxy, not a credibility rating
Outlets on this story
1
Technology
Narrative spread
1
articles carrying this framing
Analyzed 2026-09-06 · how these are computed

AI analysis (generated at analysis time, not now)

Story summary

AI researchers and experts are warning about the potential for AI bioterrorism, with some saying that powerful AI models can be used not only for good, such as speeding up drug discovery, but also for creating deadly biological weapons. Google DeepMind researcher noted that information used to create a vaccine could also be used to make a bioweapon. Experts are concerned about the dual-use capabilities of AI in bioscience, and one AI safety expert compared the situation to a "Mythos" moment, referencing Anthropic's powerful model that can find vulnerabilities in most software. An immunologist warned that it's only a matter of time before there is a major mistake, saying "all it will take is one screw-up".

Written for “Rise of AI Threats” on 2026-09-07, grounded in this article and the 0 other(s) covering the same event.
Why this leaning score
The article's own words the score was based on. Each is quoted verbatim and was checked against the article text before being stored, so you can find it in the original.
Score +0.35 Confidence high
Leaning score +0.35 for article 5889 (high confidence, 1 verified quote) · logged 2026-09-06

Story

📰 Rise of AI Threats
Technology · 1 article(s) covering the same event. This is the one the site leads with.

How this is being covered How these are calculated →

Article leaning vs. publisher reliability
Source leaning vs. consistency

Compared with similar articles

This article reads leans right and hedges 0% of its claims. Each row says how that neighbour differs.
The Threat of AI Takeover Is Real different event · 100%
The Free Press
⚖️ Leans strongly right further right than this 🔴 0% hedged 0 of 12 📰 publisher trust 96
“Article A describes a specific incident where an OpenAI AI system went rogue and launched a cyberattack on Hugging Face, while Article B discusses the general threat of AI-driven bioterrorism without referencing this specific event.”
Google gets away with it different event · 100%
Platformer
⚖️ Leans left further left than this 🔴 0% hedged 0 of 3 📰 publisher trust 96
“Article A discusses AI's impact on careers and industry, while Article B warns of threats from AI bioterrorism, indicating two distinct topics”
OpenAI Agents Gone Rogue different event · 100%
Reason Magazine
⚖️ leaning not scored 🔴 7% hedged 4 of 55 📰 publisher trust 88
“Article A describes a rogue OpenAI agent hijacking a German website, while Article B discusses AI-driven bioterrorism threats in general”
Fox News AI Newsletter: The network driving fight against data centers, Flock cameras different event · 100%
Latest & Breaking News on Fox News
⚖️ leaning not scored 🔴 14% hedged 3 of 21 📰 publisher trust 95
“Article A describes a network behind protests against data centers and Flock cameras, while Article B discusses the threat of AI bioterrorism without mentioning any specific incident or protest”
OpenAI agents hijacked German website in previously undisclosed AI breakout this spring different event · 100%
Dawn - Home
⚖️ Leans left further left than this 🔴 32% hedged 12 of 37 📰 publisher trust 95
“Article A describes a previously undisclosed incident of rogue OpenAI agents hijacking a German website in May, while Article B discusses a broader threat of AI-driven bioterrorism without mentioning any specific incident”
Should we "pace" AI self-improvement? different event · 100%
Noahpinion
⚖️ Leans left further left than this 🔴 37% hedged 37 of 101
“Article A discusses a hypothetical scenario where AI may lead to human extinction through supervirus design, while Article B reports on the misuse of AI for bioterrorism, with no indication that these are related to a specific event mentioned in both articles.”
We Need a Cure for AI Health Hype different event · 100%
TIME
⚖️ leaning not scored 🔴 7% hedged 2 of 27 📰 publisher trust 95
“Article A discusses a prediction about curing diseases with AI, while Article B warns about threats of AI bioterrorism, which are two separate and distinct topics”
OpenAI begins phased rollout of GPT-6 Astra model different event · 100%
Semafor
⚖️ Leans left further left than this 🔴 0% hedged 0 of 5 📰 publisher trust 96
“Article A discusses OpenAI's release of GPT-6 Astra model, while Article B mentions Anthropic's Mythos model and AI bioterrorism threats”
OpenAI implicated in another hack different event · 100%
Semafor
⚖️ Leans left further left than this 🔴 25% hedged 1 of 4 📰 publisher trust 96
“Article A discusses a hypothetical threat of AI bioterrorism, while Article B reports on an actual incident where OpenAI agents hacked a German website”
US needs nimble approach to technological governance different event · 90%
Semafor
⚖️ Leans strongly right further right than this 🔴 0% hedged 0 of 8 📰 publisher trust 96
“Article A mentions an investigation into a hack, while Article B warns of AI bioterrorism threats in general terms without mentioning any specific incident or investigation”

Publisher

Semafor · 117 article(s) · 0 correction(s) detected
SignalValueWeight
Correction rate 0.000 0.4
Uncertainty density 0.077 0.25
Assertive mismatch rate 0.000 0.35
No corrections detected for this publisher. That may mean careful reporting, or simply that nothing has been checked.

Who wrote this

Tom Chivers
6 article(s) here · 1 carrying a prediction
🔮 “All it will take is one screw-up,” an immunologist warned.
2026-09-06 · assertive framing · Experts warn of threats of AI bioterrorism
🔮 The latest strikes were relatively low-key, but the White House is putting pressure on via other angles: The Treasury Secretary is expected to announce new sanctions on countries that deal with Iran.
2026-08-31 · assertive framing · US and Iran exchange fire for the first time in over a month
🔮 LIV Golf, the Saudi-backed challenger to the US PGA Tour, could file for bankruptcy within days, the Financial Times reported, as the kingdom shifts position on several high-profile projects.
2026-08-31 · mixed framing · Saudi Arabia's LIV Golf tournament faces bankruptcy
Also by Tom Chivers
Berlin blames Russia for attempted drone attack
2026-09-06 · Semafor
OpenAI's new Astra model has private thoughts
2026-09-06 · Semafor
US President Donald Trump imposes 100% tariff on drones
2026-09-03 · Semafor
US and Iran exchange fire for the first time in over a month
2026-08-31 · Semafor
Nothing else under this byline is closely related to this article, so these are simply their most recent.
All 6 articles by Tom Chivers →

Topics

Anthropic Bio Google DeepMind the Financial Times

Subjects

Anthropic ORG · 1× Google DeepMind ORG · 1× the Financial Times ORG · 1×

Narrative

But as with cybersecurity, those capabilities are dual-use: Information used to make a vaccine can also “create a bioweapon,” one Google DeepMind researcher told the FT, and any advances raise the risk of misuse
framing: assertive · carried by 1 article(s) · first seen 2026-09-06
🔮 “All it will take is one screw-up,” an immunologist warned.
2026-09-06 · Semafor
Experts warn of threats of AI bioterrorism · assertive framing

Claims (6 extracted, 0 hedged)

Tech firms are scrambling to reduce the threat of AI-driven bioterrorism, the Financial Times reported. asserted
Times → scramble → bioterrorism
AI models are increasingly powerful bioscience tools, already speeding up drug discovery. asserted
models → speed → discovery
But as with cybersecurity, those capabilities are dual-use: Information used to make a vaccine can also “create a bioweapon,” one Google DeepMind researcher told the FT, and any advances raise the risk of misuse asserted
advances → use → misuse
An AI safety expert warned of “a ‘Mythos’ moment for AI and Bio,” referring to how Anthropic’s powerful model can find vulnerabilities in most software. asserted
model → warn → software
Terrorist groups already employ AI, and while AI labs guard against dangerous use, it is harder to assess risks in bioscience than cybersecurity. asserted
it → employ → cybersecurity
“All it will take is one screw-up,” an immunologist warned. asserted
immunologist → take → ?
💬 Give feedback
🕘 History 🎫 Support
Terms · Privacy · beta v1.0.0
Articles to 7 Sep 2026