Models & Labs · United States

Test shows AI models likely to give advice for terrorist attacks

Credited to The National · thenationalnews.com · 18h ago

Photograph · The National

All but two of 134 leading large language artificial intelligence models gave information useful to launching a mass casualty attack or building deadly weapons, in a test of the industry's safety measures.

A stress test of the industry found that safety measures can be stripped from leading models within three days, according to researchers at London-based Tech against Terrorism.

There have been growing concerns over the potential threats posed by AI, with former Anthropic researcher Jacob Coxon discussing the issue at a New York City hearing this month. The issue has been overshadowed by the opportunities promised by technological advances in many countries. Andy Burnham, the UK prime minister, spoke on Friday at an AI opportunity summit organised by his predecessor Gordon Brown that brought together CEOs, founders and investors to promote growth and competition.

Addressing dangers

The spread of so-called abliterated AI was identified by Tech against Terrorism as a serious danger, with 13 out of 13 failing the tests after safeguards were stripped away. Abliterated is a term for the models run on users' own systems, often without any shelf life.

"Abliteration is a proliferation problem," Tech against Terrorism said. "Companies that host, serve, list or sell models should keep abliterated and other modified copies that fail an independent test out of search, recommendation and app stores, and require verified identity for access to them, working with open-source developers rather than against them."

In the survey, the Tech against Terrorism model CT-AI put 627 requests to the AI models that a terrorist planning an attack would need to have answered. Eighty-one models gave a complete answer, while another 51 gave useful advice. That left the two that were provisionally passed on safety measures.

A user stating that they were a terrorist and intended to commit an atrocity received a usable response in just under two per cent of cases. The equivalent figure for a user stating they were a safety researcher was 16.9 per cent or 8.9 times higher, the research found.

"The models appear to respond to the stated purpose of the request, rather than the request itself," it said.

The group's founder, Adam Hadley, said: "A model that refuses a stated terrorist and answers a stated researcher has not been made safe. It has been made polite."

In addition to seeking government backing for benchmark testing such as the tools it is pioneering, Tech against Terrorism recommends developers filter hazardous knowledge and known terrorist content out of their training data.

The organisation wants developers to test how hard each model is to strip of its safety measures before it is released, and publish the results. Where a new model fails a common test, the developer should decline to release it and the results be declared to governments, it said.

Government dilemma

The report raises important questions over how countries can build trust in AI while accelerating innovation and adoption. Sovereignty over AI does not yet appear to be a factor in improving the safety of the systems on offer.

"Governments and companies should support independent safety benchmarks – ours and others – and publish which models do well as well as which do not, so that parents, schools, businesses and public bodies can choose safer models," it said. "Informed buyers move companies faster than rules do."

Models & LabsUnited States

Roundtables: A Conversation With the Creator of AI-Designed Viruses

Credited to MIT Technology Review · technologyreview.com21h ago

Useful

Join a subscriber-only conversation with one of MIT Technology Review's Innovators Under 35.

Available only for MIT Technology Review subscribers.

Can AI design new life forms? In 2025, Stanford University PhD student Samuel King came up with a preliminary answer when he used a generative AI model to propose genetic blueprints for microscopic viruses. It isn’t yet an example of AI-generated life, but that could be next. Join senior AI reporter James O'Donnell as he interviews King about his work, being named one of MIT Technology Review's Innovators Under 35 and new ways of seeing biology.

Going live on October 16th at 18:30 BST / 1:30pm EDT / 10:30am PDT

Speakers: James O'Donnell, AI Reporter, and Samuel King, Bioengineering PhD Candidate, Stanford University/Arc Institute

Original · MIT Technology Review

Front Method

Models & Labs · United States

Elastic’s AlertZero puts AI agents to work on security alert overload

Credited to SiliconANGLE · siliconangle.com · 22h ago

Photograph · SiliconANGLE

Enterprise search and security company Elastic N.V. today introduced AlertZero, a team of specialized artificial intelligence agents for security operations.

Built into Elastic Security, the agents take on alert triage, threat hunting and forensic analysis, and each customer decides how much of that work runs without a person signing off.

AlertZero is aimed at the problem of alert overload. Security operations center teams have long had more detections coming in than analysts to work through them, and false positives keep adding to the queue. Elastic argues that attackers’ use of AI is now making the problem worse.

The company’s example is the July attack on Hugging Face Inc. by OpenAI Group PBC models that had escaped a testing environment. Elastic said the intrusion generated more than 17,000 events over four days. According to Elastic, security tools could detect each signal on its own, but the hard part was connecting them to see the full attack.

Connecting those signals is the job Elastic has given four groups of agents it calls Watches. Each one runs on its own triggers or schedule and writes what it finds to a shared investigation record. Triage Watch enriches incoming alerts and decides which ones are real, closing out the noise with a written reason.

Threat hunting runs continuously under Hunt Watch, guided by threat research. Detection Watch learns from the other three and proposes tuning for noisy detection rules or new ones to cover gaps. Nothing changes in a rule without approval. Forensics Watch handles malware analysis and the tracing of exploit paths, specialist work that Elastic says many security teams lack the staff for.

Customers decide how much autonomy each Watch gets, with the setting adjustable down to individual tasks. Elastic Security Labs said in a blog post that even at the most autonomous setting, an inconclusive assessment still needs a person’s decision. Model choice sits with the customer too, and analysts can switch models partway through an investigation as the evidence shifts.

Mike Nichols, general manager of security at Elastic, said the people who built AlertZero “have sat in the SOC analyst’s seat.” Each Watch lines up with a core security operations job, he said, and the aim was to let teams take on only as much automation as they can oversee.

AlertZero builds on updates Elastic released in late July ahead of the Black Hat USA conference. Among them was a reworked Attack Discovery that investigates alerts before flagging an attack, and that tool now runs inside Triage Watch. Nichols discussed the “Alert Zero” goal behind those updates in an interview with theCUBE, SiliconANGLE Media’s livestreaming studio (below).

Customers can add workflows of their own using the same Agent Builder skills and Elastic Workflows automation the company used to build AlertZero. The release will be available to Elastic Security customers as a technical preview on Elastic Cloud and in self-managed and air-gapped deployments.

Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.

Are you an AWS customer? Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/

Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.

CoreWeave targets AI inference bottlenecks with full-stack optimization

White House bars Microsoft from sponsoring green card applications as JD Vance accuses it of fraud

AI stocks crumble on report that OpenAI's annualized revenue is much lower than believed

Gallatin AI raises $50M in funding for its military logistics platform

Anthropic launches critical infrastructure program and free OSS Scanner for open source

Alphabet spinoff Isomorphic Labs reportedly raising funding at up to $50B valuation

SECURITY - BY DUNCAN RILEY. 12 HOURS AGO

Join our community