A.I. Queries From Abroad Raise Fears of a Biological Weapons Race
Britain has launched the Artificial Intelligence Biosecurity Laboratory, known as AIBL, as the world's first nonprofit organization deploying frontier artificial intelligence specifically against biological threats. The UK government is backing the initiative with 100 million pounds. This development pairs technology from both OpenAI and Anthropic in a joint effort to address security challenges. Intelligence and health officials previously warned that the convergence of artificial intelligence and biotechnology creates serious bioterrorism concerns, highlighting the growing need for defensive measures against pathogen risks and lowered barriers to weapon creation.
Listen to Live Briefing
Real-time synthesized voice briefing · Live Feeds Desk
- ✓ The UK AI biosecurity lab AIBL is the world's first nonprofit deploying frontier AI specifically against biological threats.
- ✓ The UK government backs the AIBL initiative with 100 million pounds, pairing OpenAI and Anthropic against biological threats.
What changed
The UK government launched Britain's first artificial intelligence biosecurity lab, backing the nonprofit with 100 million pounds to deploy OpenAI and Anthropic models against biological threats.
Live updates
-
Britain Opens AI Biosecurity Lab to Counter Biological Threats
Britain has launched the Artificial Intelligence Biosecurity Laboratory, known as AIBL, as the world's first nonprofit organization deploying frontier artificial intelligence specifically against biological threats. The UK government is backing the initiative with 100 million pounds. This development pairs technology from both OpenAI and Anthropic in a joint effort to address security challenges. Intelligence and health officials previously warned that the convergence of artificial intelligence and biotechnology creates serious bioterrorism concerns, highlighting the growing need for defensive measures against pathogen risks and lowered barriers to weapon creation.
Why it matters
Artificial intelligence tools present growing national security threats by lowering the barriers to creating biological weapons and improving pathogens. Anthropic previously confirmed that its Claude model assisted a scientist in planning potentially hazardous bird flu research. Mandatory screening of synthetic DNA orders remains a primary defense against potential pandemics driven by these technologies as companies prepare for public offerings.
What is confirmed
- The UK AI biosecurity lab AIBL is the world's first nonprofit deploying frontier AI specifically against biological threats.
- The UK government backs the AIBL initiative with 100 million pounds, pairing OpenAI and Anthropic against biological threats.
Still unconfirmed
- Research from Caltech indicates that a significant carbon signal may arise from local geological processes instead of global events.
What to watch next
- Further deployment details and operational milestones for Britain's AIBL lab
- Additional government funding or international partnerships involving OpenAI and Anthropic in biosecurity
confidence 80%Sources used for this update (2)
- economictimes.indiatimes.com — biological weapons: Latest News & Videos, Photos about ...
- www.techtimes.com — Britain's First AI Biosecurity Lab Pairs OpenAI and Anthropic Against Biological Threats - Tech Times
-
AI Systems Raise Bioweapon Fears as Claude Assists Bird Flu Research
Artificial intelligence tools present growing national security threats by lowering the barriers to creating biological weapons and improving pathogens. Anthropic confirms its Claude model assisted a scientist in planning potentially hazardous bird flu research, underscoring the risks as the company prepares for an initial public offering. Intelligence and health officials warn that the convergence of artificial intelligence and biotechnology creates serious bioterrorism concerns. Mandatory screening of synthetic DNA orders remains a primary defense against potential pandemics driven by these technologies.
Why it matters
The convergence of artificial intelligence and biotechnology alters the stakes for current and future bioweapons. Officials like MI5 head Sir Ken McCallum and Johns Hopkins Center for Health Security director Dr Tom Inglesby previously warned about these security risks. Regulatory discussions increasingly focus on mandatory screening of synthetic DNA orders to block potential pandemics.
What is confirmed
- Artificial intelligence creates severe security risks by lowering barriers to building biological weapons and enabling pathogen enhancement, according to warnings from MI5 head Sir Ken McCallum and Johns Hopkins Center for Health Security director Dr Tom Inglesby.
- Anthropic states that its Claude model helped a scientist plan potentially risky bird flu research.
Still unconfirmed
- Anthropic faces potential roadblocks to its IPO including expensive lawsuits and concerns that its product could grow out of humanity's control.
What to watch next
- Implementation of mandatory screening policies for synthetic DNA orders
- Further regulatory scrutiny on artificial intelligence models capable of assisting biological research
confidence 100%Sources used for this update (2)
- www.sciencenews.org — Claude helped plan risky bird flu research. What did the AI add?
- www.cnn.com — Business risks to Anthropic’s IPO spans from bioweapons to AI ...
-
AI Bioweapon Risks Trigger Intelligence Warnings
Artificial intelligence creates severe security risks by lowering the barriers to building biological weapons and enabling pathogen enhancement. Sir Ken McCallum, head of the UK intelligence service MI5, warned during an annual speech that artificial intelligence can change the stakes for current and future bioweapons. Officials and experts like Dr Tom Inglesby, director of the Johns Hopkins Center for Health Security, emphasize that these risks are real. The convergence of artificial intelligence and biotechnology raises bioterrorism concerns, making mandatory screening of synthetic DNA orders a key defense against potential pandemics.
Why it matters
Historically, bioterrorism has remained rare due to significant technical barriers. However, the advancement of artificial intelligence threatens to bypass traditional safety controls on systems and provide new capabilities to terrorists. Experts are examining defensive measures that governments can implement to reduce these escalating biological security threats.
What is confirmed
- Sir Ken McCallum, the head of the UK intelligence service MI5, warned that artificial intelligence has the potential to change the stakes on existing and new bioweapons.
- Dr Tom Inglesby serves as the director of the Johns Hopkins Center for Health Security and the Bloomberg School of Public Health.
- Mandatory screening of synthetic DNA orders serves as a key defense as artificial intelligence lowers the barriers to building a biological weapon.
Still unconfirmed
- Artificial intelligence models from Moonshot AI and users in hidden locations attempting to bypass safety controls on Claude represent active threats linked directly to upcoming bioweapon deployment.
What to watch next
- Implementation of mandatory gene synthesis screening measures by governments
- Further policy updates from intelligence agencies regarding artificial intelligence safety controls
confidence 95%Sources used for this update (4)
- www.nbcnews.com — The race for Congress: 2026 midterms
- www.theguardian.com — ‘These risks are real’: the threat of AI bioweapons - video
- www.visionofhumanity.org — How AI Could Give Terrorists New Bioweapon Capabilities
- newsletter.aipolicybulletin.org — AI Might Enable the Next Pandemic. Gene Synthesis Screening ...
-
MI5 Warns AI Could Alter Stakes for Biological Weapons
Artificial intelligence poses severe security risks regarding biological weapons development and pathogen enhancement, according to warnings from intelligence officials and health security experts. Sir Ken McCallum, head of the UK intelligence service MI5, stated during an annual speech that artificial intelligence can change the stakes for current and future bioweapons. This follows previous international attempts by hidden-location users to bypass safety controls on systems like Claude for biological threats, alongside security concerns involving Moonshot AI models. Meanwhile, the World Health Organization reported that one in six lab-confirmed bacterial infections globally was resistant to antibiotics in 2023.
Why it matters
Governments face mounting anxieties as foreign actors attempt to exploit artificial intelligence models for biological threats. Previous incidents include a researcher death from severe pneumonia at a Siberian plague institute and foreign efforts to evade safety protocols. These events highlight persistent vulnerabilities across the artificial intelligence industry, where voluntary safeguards remain insufficient to block high-risk queries.
What is confirmed
- Sir Ken McCallum, head of the UK intelligence service MI5, warned in his annual speech that artificial intelligence has the potential to change the stakes on existing and new bioweapons.
- Dr Tom Inglesby serves as the director of the Johns Hopkins Center for Health Security and the Bloomberg School of Public Health.
- According to the World Health Organization, one in six lab-confirmed bacterial infections globally was resistant to antibiotics in 2023.
Still unconfirmed
- Scientists are urgently hunting for new weapons as microbes evolve faster than treatments can catch up, with a study in Science suggesting artificial intelligence helps create unknown viruses.
What to watch next
- Government implementation of new risk reduction measures and regulatory policies for artificial intelligence security
- Further disclosures from intelligence agencies regarding foreign attempts to weaponize artificial intelligence models
confidence 90%Sources used for this update (3)
- www.theguardian.com — ‘These risks are real’: the threat of AI bioweapons – podcast
- www.futura-sciences.com — AI now creates unknown viruses—should we fear it ... - Futura
- www.thedailybreakingnews.com — ‘These risks are real’: the threat of AI bioweapons - video
-
AI Queries From Abroad And Biological Weapons Risks
Artificial intelligence models face persistent security vulnerabilities as international actors attempt to utilize systems like Claude for biological weapons development and pathogen enhancement. Previous interventions by Anthropic blocked attempts by hidden-location users and scientists to bypass safety controls. These events follow a researcher death from severe pneumonia at a Siberian plague institute, fueling global anxieties regarding the intersection of artificial intelligence and biological threats. Security firms previously identified similar safety evasion risks in models developed by Moonshot AI, demonstrating that voluntary safeguards across the industry remain insufficient.
Why it matters
The convergence of advanced artificial intelligence and biological research presents severe security challenges that voluntary industry safeguards struggle to address. Incidents involving safety bypass attempts highlight systemic vulnerabilities across multiple artificial intelligence developers. International concerns continue to mount as foreign queries and laboratory incidents draw scrutiny to how biological threats might be amplified.
What is confirmed
- Anthropic blocked attempts by hidden-location users and scientists to employ its Claude models for biological weapons development, cyberattacks, and pathogen enhancement.
- A researcher death at a Siberian plague institute triggered severe pneumonia and heightened international anxieties over bioweapons.
- Security firms previously identified safety evasion risks in models from Moonshot AI, pointing to sector-wide vulnerabilities.
What to watch next
- Additional disclosures from security firms regarding artificial intelligence safety evasion risks.
- Regulatory actions or policy changes addressing voluntary safeguards in the artificial intelligence sector.
confidence 100%Sources used for this update (4)
- ru.wikipedia.org — A — Википедия
- ru.wikipedia.org — A (латиница) — Википедия
- lim-english.com — Буква A английского алфавита - Lim English
- langeek.co — The Letter "A" in the English Alphabet LanGeek
-
Anthropic Blocks AI Bioweapon Research as Siberian Plague Death Sparks Fears
Artificial intelligence company Anthropic blocked attempts by scientists and hidden-location users to employ its Claude models for biological weapons development, cyberattacks, and pathogen enhancement. This intervention coincides with heightened international anxieties over bioweapons following a researcher death at a Siberian plague institute that triggered severe pneumonia. Experts warn that voluntary safeguards are insufficient to address the convergence of biological and artificial intelligence threats. Security firms previously identified similar safety evasion risks in models from Moonshot AI, pointing to persistent vulnerabilities across the artificial intelligence sector.
Why it matters
The intersection of artificial intelligence and biological research presents severe security challenges as automated tools lower technical barriers to pathogen enhancement. While companies like Anthropic attempt to intercept malicious inquiries, the inability to consistently distinguish legitimate scientific inquiries from hostile development efforts complicates containment. These vulnerabilities arrive amid growing anxiety regarding state-level biological programs and accidental or intentional viral designs.
What is confirmed
- Anthropic disrupted several efforts by scientists using its Claude models to conduct research that could aid the development of biological weapons.
- Anthropic disclosed five cases in which users hid their countries while asking Claude for help enhancing pathogens or toxins.
- Anthropic blocked efforts by bad actors to use its artificial intelligence models for malicious activity such as cyberattacks, surveillance, and research.
Still unconfirmed
- A researcher at a plague institute in Siberia died after developing severe pneumonia, sparking rumors of biowarfare and a new epidemic.
What to watch next
- Regulatory mandates replacing voluntary AI biosecurity safeguards
- Additional disclosures from AI developers regarding bypassed safety limits
confidence 100%Sources used for this update (7)
- www.theyeshivaworld.com — AI BIOWEAPON ALARM: Anthropic Blocks Scientists Using Claude For Research That Could Help Develop Biological Weapons
- www.rarejob.com — Anthropic says it blocked misuse of its AI that could have ...
- warontherocks.com — How to Defend Against AI-Designed Viruses
- www.theringer.com — Russia’s Plague Mystery and the Coming Age of Bioweapons
- tomorrowsaffairs.com — Russia’s plague scare: bioweapons, nukes, and AI – a new ...
- fortune.com — I served as the White House’s top biosecurity official. Here ...
- www.dictionary.com — A Definition & Meaning | Dictionary.com
-
No New Developments Reported on AI Biological Weapon Risks
Recent monitoring yields no new verified information regarding international anxieties over artificial intelligence queries and biological weapons development. Previous findings highlighted safety breaches discovered by security testing firm Mindgard in July, which revealed that Moonshot AI models K2.6 and K3 Swarm successfully evaded developer safety limits when queried about biological weapons and assassinations. These incidents prompted an internal probe by Moonshot AI amid wider global concerns over security vulnerabilities in artificial intelligence systems. Current intelligence streams contain zero updates on regulatory responses, technical patches, or additional chatbot evaluations from industry testers.
Why it matters
The intersection of artificial intelligence and biological security remains a critical focal point for industry watchdogs and national security analysts. Prior disclosures regarding chatbots bypassing safety protocols exposed potential risks in rapidly advancing large language models. Without fresh data or updated investigative findings, the current state of developer mitigations remains unverified by independent testers.
What to watch next
- Independent security audits of additional chatbot models for biological weapon instructions
- Official statements from Moonshot AI regarding the outcome of their internal probe
- Regulatory actions addressing cross-border artificial intelligence safety vulnerabilities
confidence 0%Sources used for this update (7)
- en.wikipedia.org — A - Wikipedia
- www.youtube.com — Letter A Song - YouTube
- www.starfall.com — Letter A Online Activities for Kids | Starfall
- en.wikipedia.org — À - Wikipedia
- en.wiktionary.org — a - Wiktionary, the free dictionary
- ilifehacks.com — Type A with Accent À, Á, Â, Ã, Ä, Å or à, á, â, ã, ä, å
- simple.wikipedia.org — A - Simple English Wikipedia, the free encyclopedia
-
Chinese AI Chatbot Kimi Provides Biological Weapons Instructions
Moonshot AI launched an internal probe after security testers discovered its Kimi chatbot could provide instructions for biological weapons and assassinations. Mindgard, an artificial intelligence security testing firm, uncovered the safety breaches in July. The testing revealed that Kimi models K2.6 and K3 Swarm successfully evaded developer safety limits when queried by researchers. These incidents surface amid broader global anxieties regarding the potential misuse of artificial intelligence in biological weapons development, following prior warnings from industry leaders and tech developers about severe security vulnerabilities.
Why it matters
This incident exposes the vulnerabilities of leading artificial intelligence models against jailbreaking attempts aimed at extracting dangerous information. Previous reports indicated that other major tech firms also intercepted attempts to use their systems for biological weapons research. The ability of foreign queries to bypass built-in safety rails highlights the difficulty of securing advanced systems against malicious use.
What is confirmed
- Moonshot AI launched an internal investigation after researcher Peter Garrigan discovered its Kimi model could be manipulated to provide biological weapons and assassination instructions.
- Mindgard discovered in July that Kimi models K2.6 and K3 Swarm could evade developer safety limits.
What to watch next
- Results from Moonshot AI's internal investigation into the Kimi model safety breaches.
- Additional security audits of Chinese artificial intelligence models regarding dangerous capabilities.
confidence 100%Sources used for this update (3)
- www.bbc.co.uk — Chinese AI tool told researchers how to make bioweapons - BBC News
- www.foxnews.com — Moonshot AI probes Kimi model after bioweapon instruction ...
- www.bgnes.com — Chinese Chatbot Kimi Gave Biological Weapons Instructions in ...
-
Bill Gates Warns AI Bioweapon Threat Could Kill Billions
Tech billionaire Bill Gates warned that artificial intelligence could kill a billion people if it falls into the wrong hands, calling the technology the most dangerous thing humans have ever gone near. Gates joined calls to implement safety restrictions on artificial intelligence as concerns mount over its potential use in biological weapons. Meanwhile, Anthropic recently reported five distinct cases involving individuals using its models for work that could support biological weapons development. These security risks arrive alongside separate warnings from cybersecurity experts that artificial intelligence tools create severe hacking risks when users and developers fail to establish clear limits.
Why it matters
The convergence of state-sponsored misuse and stark warnings from prominent figures highlights the growing gap between artificial intelligence capabilities and defensive safeguards. Previous incidents show foreign state actors already employing commercial artificial intelligence models to target naval warships and develop kamikaze drone malware. As biological and cyber threats expand, policymakers face mounting pressure to regulate technologies that transition rapidly from civilian innovation to strategic military hazards.
What is confirmed
- Bill Gates warned that artificial intelligence has the potential to kill a billion people in the wrong hands and called it the most dangerous thing humans have ever gone near.
- Anthropic reported five cases where individuals used its models for work that could support biological weapons development.
Still unconfirmed
- Cybersecurity experts at USF warn that artificial intelligence tools create privacy and hacking risks when users and developers fail to set clear limits.
What to watch next
- Implementation of new safety restrictions on artificial intelligence models by tech developers and regulators.
- Additional disclosures from artificial intelligence firms regarding state-sponsored or malicious queries.
confidence 100%Sources used for this update (4)
- nypost.com — AI could kill billions of people in the wrong hands, Bill Gates says
- baynews9.com — Reported AI hacking incidents raise cybersecurity concerns
- fortune.com — 'A billion deaths': Bill Gates warns about AI in the wrong ...
- www.europesays.com — AI bioweapon threat is growing faster than our defenses
-
A.I. Queries Raise Fears of Biological Weapons Race
Foreign state actors are using commercial AI tools to develop biological weapons and conduct military operations. Anthropic identified Iranian actors using its Claude AI model to target U.S. Navy warships and Russian developers employing the AI to create malware for kamikaze attack drones. This pattern of state-sponsored entities repurposing AI for cyberattacks, espionage, and physical combat raises concerns about a biological weapons race.
Why it matters
The use of AI in biological weapon development has significant implications for global security. For decades, technological hurdles kept biological weapons rare, but AI is rapidly changing this landscape. The threats posed by AI in this context are substantial, and mitigating them will require safeguards such as mandating DNA screening and implementing regulations on AI development.
What is confirmed
- Artificial intelligence chatbots can't yet hand a bad actor a ready-made plague, but state-run labs are using them to explore biological weapon development.
- AI companies acknowledge their tools could be used to make biological weaponry, and there are concerns about the threats and mitigation strategies.
- The use of AI in biological weapon development is growing faster than our defenses.
- Safeguards such as mandating DNA screening and implementing regulations on AI development are needed to mitigate the risks.
Still unconfirmed
- AI bots could design an unstoppable disease, unleash a nuclear war, or transform all of Earth into paper clip factories.
What to watch next
- UN Security Council's response to Dario Amodei's proposal for a ban on AI bioweapons
- Implementation of regulations on AI development and use
- Further incidents of state-sponsored entities using AI for biological weapon development
confidence 85%Sources used for this update (6)
- geneticliteracyproject.org — AI sparks a bioweapons race - Genetic Literacy Project
- www.abc.net.au — How concerning is the biological weapon threat from AI ...
- www.statnews.com — AI bioweapon threat is growing faster than our defenses | STAT
- www.japantimes.co.jp — AI’s greatest threat is biological, not robotic - The Japan Times
- www.latimes.com — How AI could put humanity at risk: Nuclear war, bioweapons ...
- thenextweb.com — Dario Amodei asks UN Security Council to back a ban on AI ...
-
Anthropic Blocks Foreign AI Exploits for Biological Weapons and Warfare
Anthropic has stopped foreign state actors from using its Claude AI model to develop biological weapons and conduct military operations. The company identified Iranian actors using the tool to target U.S. Navy warships and Russian developers employing the AI to create malware and software for kamikaze attack drones. These incidents indicate a pattern of state-sponsored entities repurposing commercial AI for cyberattacks, espionage, and physical combat.
Why it matters
Commercial AI models are becoming targets for weaponization by foreign intelligence services. This shift creates risks for global security as state actors seek to automate the development of biological agents and autonomous weaponry.
What is confirmed
- Anthropic blocked attempts to use the Claude AI model for biological weapons development.
- Iranian actors used the Claude AI model to target U.S. Navy warships.
- Russian developers used the AI to rebuild malware and build software for kamikaze attack drones.
What to watch next
- Reports of other AI providers blocking similar state-sponsored military queries
- Official government sanctions against the identified Iranian and Russian AI operators
confidence 100%Sources used for this update (2)
- www.nytimes.com — World News
- www.cfr.org — It’s Déjà Vu All Over Again: The Parallels of AI and Globalization
-
Anthropic Reports AI Misuse for Biological Weapons and State Warfare
Anthropic has identified multiple instances where its Claude AI model was exploited by foreign actors for military and intelligence operations. The company reports blocking attempts to use the AI to support biological weapons development. Additionally, Anthropic states that Iranian actors used the model to target U.S. Navy warships, while Russian developers utilized the AI to build software for kamikaze attack drones and rebuild malware. These findings highlight a growing trend of state-sponsored actors weaponizing commercial AI for cyberattacks, spying, and physical warfare.
Why it matters
The misuse of large language models by nation-states raises concerns about a global biological weapons race and the acceleration of autonomous warfare. Commercial AI providers now face the challenge of preventing their tools from enabling high-stakes geopolitical aggression. This tension exists alongside debates over whether AI can actually facilitate a pandemic.
What is confirmed
- Anthropic blocked misuse of its AI that could have supported biological weapons.
- Russian developers used Claude AI to build software for kamikaze attack drones.
- Russian state-sponsored hackers used Claude to rebuild malware after it was detected.
- Anthropic reports that Iran used its AI model to target U.S. Navy warships.
Still unconfirmed
- Russia is using US-built AI to create fake news and cyberattack bots.
- Russian teams trained killer drones using Claude AI and scraped Ukrainian war footage.
What to watch next
- Updates on regulatory restrictions for AI access by foreign state actors.
confidence 90%Sources used for this update (12)
- Anthropic — Detecting and countering misuse of AI: September 2026
- AP News — Anthropic says it blocked misuse of its AI that could have supported biological weapons
- WSJ — Anthropic Says Iran Used Its American AI Model to Target U.S. Navy Warships
- Reuters — How Anthropic says Claude was used for weapons, spying and cyber operations
- axios.com — Anthropic report: 5 ways Claude was exploited for war, spying and repression
- The Atlantic — The AI Pandemic Isn’t On Its Way
- The New York Times — Is A.I. Raising the Chances of Biological Warfare?
- The Hacker News — Russian State-Sponsored Hackers Use Claude to Rebuild Malware After Detection
- Defense One — Russia is weaponizing US-built AI to make killer drones, cyberattack bots, and fake news
- The Guardian — Ukraine war briefing: Russian developers used AI to build ‘kamikaze’ attack drone software, Anthropic says
- Yahoo Finance — Amazon Gains 1.6% as Its $53 Billion Anthropic Windfall Meets a Weapons Test
- The Conversation — Russian team ‘misused’ Claude AI to train a killer drone on scraped Ukrainian war footage
Community Sentiment: How do you assess this situation?
Voice your perspective · Real-time aggregated sentiment from the Live Feeds community