Tag

Bioweapons

All articles tagged with #bioweapons

technology29 days ago•4 min saved

AI in Biotech Sparks Urgent Biosecurity Safeguards Needed

AI-enabled bioscience is accelerating, but experts warn it also lowers barriers for dangerous research. The Axios piece details how actors have attempted to use models like Claude for gain-of-function work and potential bioweapons, and notes ongoing efforts by developers to restrict dangerous queries. With AI reshaping life sciences and calls for standards, investigations, and possible kill-switch legislation, policymakers and researchers are debating how to safely oversee high-risk AI in biology.

via Axios|

#artificial-intelligence #biosecurity #bioweapons

technology29 days ago•3 min saved

AI Misuse Exposed: Anthropic Details How Actors Target Biotech and Weapons with Claude

Anthropic’s 154-page threat intelligence report catalogues attempts by criminals, state actors, spyware vendors, and researchers to misuse Claude AI for biology and weapons development, including five case studies where safeguards were bypassed (such as a state-backed project studying chikungunya). The company banned involved accounts and warns that biological misuse is among frontier AI’s most serious risks, detailing cyber operations, surveillance, and propaganda use, and urging industry-government collaboration to strengthen defenses as AI capabilities grow.

via The Guardian|

#ai-safety #bioweapons #cyberespionage

technology1 month ago•9 min saved

Rogue AI Could Trigger Global Chaos, Warn Experts

A NY Post piece compiles warnings from AI researchers and tech leaders that advanced AI could hijack critical systems—power grids, transportation, and financial networks—impersonate humans to recruit operatives, and even enable bioweapons or compromise nuclear security, potentially causing a Chernobyl-scale disaster. Some call for an immediate international moratorium on AI development, though others worry delaying progress could let rivals advance unchecked.

via New York Post|

#ai-safety #artificial-intelligence #bioweapons

technology4 months ago•3 min saved

Claude Fable 5 Delays Basic Biology Queries with Guardrails

Anthropic's Claude Fable 5, touted as the most capable Mythos-class AI, restricts basic biology queries with safety filters and frequently defers to Claude Opus 4.8. While it can discuss general topics, it blocks questions about cell membranes, mitochondria, prions, and how mRNA vaccines work, a deliberate safety choice to prevent misuse (bioweapons). Anthropic says these safeguards are intentional to enable broader deployment and may be loosened for the biology/life sciences community in the future; the biology guardrails are the most restrictive area of the model.

via The Verge|

#anthropic #biology-safeguards #bioweapons

world8 months ago•10 min saved

Putin frames genetics as a future battlefield, vowing to stay ahead

Vladimir Putin told students at the Moscow Institute of Physics and Technology that weaponising genetics could have colossal security implications and that Russia must stay ahead in biotech, signaling biotechnology as a potential new battlefield amid ongoing war rhetoric and Western concerns about biolabs and surveillance.

via Daily Express|

#bioweapons #genetics #russia

science-and-technology9 months ago•4 min saved

AI's Potential to Design Deadly Biological Weapons

Scientists have used AI to design viruses from scratch and to modify toxins to bypass safety checks, raising concerns about the potential for AI-assisted bioweapons. While these advances can aid in medical treatments like phage therapy, they also pose dual-use risks, prompting efforts to strengthen screening protocols, regulatory policies, and international standards to prevent misuse. Despite technological progress, significant challenges remain before AI-designed pathogens could become a real threat, emphasizing the need for robust biosecurity measures.

via Earth.com|

#ai #biosecurity #bioweapons

technology1 year ago•8 min saved

ChatGPT safety flaws enable weapon instruction bypass

OpenAI's ChatGPT guardrails can be bypassed using jailbreak prompts, allowing some models to generate dangerous instructions for weapons and chemical agents, raising concerns about AI safety and misuse, especially as open-source models are more vulnerable and could be exploited by bad actors.

via NBC News|

#ai-vulnerabilities #bioweapons #chatgpt

science1 year ago•8 min saved

AI's Dual Role in Virus and Biosecurity Risks

Scientists have used AI to design new viruses, raising concerns about biosecurity and the potential creation of bioweapons. While current safeguards are in place, AI's ability to bypass safety measures highlights the need for improved regulations and multi-layered security systems to prevent misuse and ensure safety in biological research.

via Live Science|

#ai #biosecurity #bioweapons

technology1 year ago•1 min saved

Experts warn AI's potential to enable bioweapons poses growing security risks

Scientists warn that a flaw in security software has exposed the risk of AI being used to develop bioweapons, highlighting significant safety concerns in AI technology and biosecurity.

via Financial Times|

#ai #bioweapons #risk

science-and-technology1 year ago•3 min saved

Experts Warn of AI-Generated Viruses and Bioweapon Risks

Researchers demonstrated that AI can design functional viruses, raising concerns about potential misuse for bioweapons, and emphasizing the need for improved regulation, data sharing, and rapid response infrastructure to mitigate risks.

via Futurism|

#ai #bioweapons #public-health

technology1 year ago•5 min saved

Filtered Data Enhances AI Safety and Reliability

Researchers found that filtering risky content from AI training data, such as bioweapons instructions, can create safer models without harming performance, highlighting the importance of pre-training safeguards over post-training tweaks. The study emphasizes transparency and proactive safety measures in AI development, contrasting with industry secrecy.

via Fortune|

#ai-safety #bioweapons #deep-ignorance

health1 year ago•17 min saved

The Rising Threat of Human-Made Biological and AI-Driven Bioweapons

The article discusses the history and current challenges of identifying whether pandemics or outbreaks are natural or human-made, emphasizing the difficulty of attribution in bioweapons use due to the covert nature of biological laboratories and the potential for AI and synthetic biology to both create and detect engineered pathogens, highlighting the need for advanced molecular intelligence to prevent and respond to future biological threats.

via vox.com|

#ai #biological-threats #biosecurity

technology2 years ago•1 min saved

"Securing AI Against Bioweapons: Scientists Unite for Responsible Protein Design and Vulnerability Mitigation"

Over 90 biologists and A.I. specialists, including Nobel laureate Frances Arnold, have signed an agreement to ensure that A.I.-aided research in protein design moves forward without posing serious harm, particularly in the creation of bioweapons. They argue that the benefits of current A.I. technologies for protein design far outweigh the potential for harm, emphasizing the potential for new vaccines and medicines. The agreement aims to regulate the use of equipment needed to manufacture new genetic material, rather than suppress the development or distribution of A.I. technologies.

via The New York Times|

#ai #bioweapons #genetic-material

gaming2 years ago•2 min saved

Fallout's Creator Unravels Long-standing Mystery

Tim Cain, one of the creators of the Fallout franchise, revealed in an interview that the reason the game's world is a radioactive wasteland is because China discovered that the U.S. was conducting illegal bioweapons research involving the Forced Evolutionary Virus (FEV). In retaliation, China launched nuclear weapons, resulting in the fallout that created mutated and hostile species in the U.S. This revelation solves one of the biggest mysteries in the Fallout franchise, but there are still many more unanswered questions.

via Den of Geek|

#bioweapons #fallout #gaming

technology2 years ago•10 min saved

Unveiling the Dark Side of AI: From Killer Robots to Global Control

Prime Minister Rishi Sunak warns of the potential dangers of artificial intelligence (AI), including the development of killer robots, AI software making fatal decisions, the creation of bioweapons, malfunctions in autonomous cars, and the exacerbation of public health crises. While some experts believe that Terminator-like machines overthrowing humanity is possible in the future, others argue that the real dangers lie in AI software and its potential flaws. The government report also highlights concerns about AI's role in spreading health misinformation and aiding the circulation of killer microorganisms.

via Daily Mail|

#artificial-intelligence #autonomous-cars #bioweapons