NAll news

All news

Friday37 items

  2. Cailian PressResearch
    AI score70

    Microsoft's AI diagnostic tool simulates a doctor's step-by-step questioning, with accuracy on complex cases about twice that of general practitioners

    Microsoft's research team published in the journal Science an AI diagnostic tool powered by GPT-5, which achieved a diagnostic accuracy of about 83% on complex cases selected from the New England Journal of Medicine, compared with about 36% for general practitioners with reference materials and about 22% without them. Through a mechanism called the "Diagnostic Coordinator," the tool simulates multiple clinical experts analyzing cases collaboratively, deciding step by step which medical history to ask about and which tests to order, and continuously updating diagnostic hypotheses. The research team's next step is to collaborate with Boston's Beth Israel Deaconess Medical Center to test emergency triage capabilities; Microsoft has not yet specified the timing or form of commercialization.

    Same story · 4 reports

    #Papers/Research#Agents#OpenAI

  3. UK AI Security Institute: BlogFirst-partyResearch
    AI score71

    UK AISI evaluates frontier AI agents' performance in multi-step cyberattack scenarios

    The UK AI Security Institute evaluated seven models released between August 2024 and February 2026 on a self-built cyber range, finding that each generation of models outperformed the previous one under a fixed token budget: in a 32-step enterprise cyberattack range, the average number of completed steps under a 10M token budget rose from 1.7 steps for GPT-4o to 9.8 steps for Opus 4.6. Raising inference compute from 10M to 100M tokens yielded gains of up to 59% with no plateau observed, and at current prices a single 100M token attempt with Opus 4.6 costs about 80 USD. In a 7-step industrial control system range, model progress was limited: under 10M tokens most models completed zero steps, and under 100M tokens Opus 4.6 averaged 1.4 steps.

    #Evaluations / benchmarks#Agents#Safety / alignment

  4. heise onlineResearch
    AI score52

    W3C discussion draft proposes using Web standards to constrain AI agents

    The W3C Strategy team published a discussion draft, "W3C and the State of the Web in the AI Era," proposing the use of Web standards to make AI agents safer, focusing on Prompt Injection protection, controllable permissions, and accessibility. The article was published on October 21 to build momentum for the TPAC meeting; it is a proposal rather than an approved standard, with the official presentation scheduled for October 29. Using travel booking as an example, the article points out that letting an agent search for flights does not equal authorizing it to pay for tickets, and recommends setting spending limits, revocable authorization, and traceable credentials for payments.

    #English text is being prepared.#Agents#Safety / alignment

  5. Phoenix TechnologyResearch
    AI score74

    Google DeepMind autonomously completes Minecraft with four Gemini 4 Argon agents

    Google DeepMind published the paper "A team of Gemini 4 Argon agents beats Minecraft": four coding agents driven by Gemini 4 Argon autonomously defeated the Ender Dragon in randomly generated Minecraft survival mode, writing 2,783 program segments throughout and taking 27 hours 14 minutes. The division of labor was not pre-written into the prompt; rather, the agents proposed it themselves about ten minutes after the game started and it gradually took shape, with role responsibilities adjusting as the process went on. The team died a cumulative 605 times, and the causes of failure were analyzed and stored in shared memory for the whole team to reuse; the whole run involved about 24,000 model calls and 1.35 billion input tokens, 91% of which came from context caching, and at discounted API prices the estimated cost to complete the game was about $520, with about 22% of game programs showing errors and the server restarting 6 times.

    Same story · 3 reports

    #Papers/Research#Agents#DeepMind

  6. Shao MengResearch
    AI score83

    METR and Redwood each sent one person to OpenAI for three weeks to evaluate monitoring of internal Agent misalignment behavior

    METR and Redwood Research each sent one person to OpenAI from July 21 to August 11, 2026, to evaluate the misalignment behavior monitoring system for its internal agents (including Codex). The public draft states that about 75–94% of agent traffic in the internal Responses API at or above GPT-5.5 is eligible for monitoring, and the internal Codex agent subset is about 99.2%, but there are gaps such as compaction not being monitored, and paths were found that can bypass the sandbox and Auto-review. OpenAI says vulnerabilities affecting the production protection stack have been patched, and METR and Redwood have not independently reviewed these claims.

    Quoted postMETR · @METR_Evals

    Two employees from METR and Redwood Research worked on premises at OpenAI for three weeks this summer in an embedded assessment pilot. This team evaluated the systems OpenAI was using to monitor for misalignment from internal agents. A brief write-up of their work is now public.

    Same story · 4 reports

    #Safety / alignment#Agents#OpenAI

  7. Shao MengResearch
    AI score58

    SkyRL extends GLM-5.3-Flash training context to 1 million tokens

    Through four types of system optimizations, SkyRL achieves a single-sequence training length of 1 million tokens for GLM-5.3-Flash on B200, with a peak of about 144.5 GiB, roughly 64 times that of the out-of-the-box Megatron baseline. The optimizations include running DSA on sparse kernels, chunking by query and sharding the indexer by TP, recomputing full-layer activations and CPU offloading layer inputs, and removing redundant fp32 upcasting. On real book data, the 512k peak is about 122 GiB, and 1M (CP=4) is about 109 GiB; the mean absolute difference in logprob between the trainer and vLLM is about 0.04–0.07.

    Quoted postavi basnet · @avigyabasnet

    Super excited to share some of the work I’ve been doing on supporting 1M token context training on SkyRL for GLM 5.3 Flash, 64x higher than the out of the box Megatron baseline! 🚀 This work was done in collaboration with the amazing folks at @LilaSciences! 🤩 Getting there took key algorithmic and systems optimizations in four main areas: 1) DeepSeek Sparse Attention: run DSA on a sparse kernel; chunk and shard its quadratic indexer. 2) Activation recompute and offloading: recompute activations in backward and offload saved inputs to CPU. 3) Context parallelism: split each sequence across GPUs, with new CP support for KDA and a fix for DSA. 4) Avoiding unnecessary precision upcasts: keep FP32 copies out of memory and off the network. For more details on our implementations and our results, check out our full technical blog here: https://t.co/Dna0CGNeJt! This work benefited from the contributions of many people along the way. Big thanks to @pcmoritz, @erictang000, @sumanthrh, @CyrusHakha, @richliaw from @anyscalecompute and Keith Truongcao, Tushar Khot, Vivek Subramanian, Semih Kara from Lila Sciences for all the help and guidance!

    #English text is being prepared.#English text is being prepared.#English text is being prepared.

  8. Rohan PaulResearch
    AI score75

    Google paper finds AI agents writing tool descriptions exaggerate capabilities

    A new Google paper finds that when AI agents write tool descriptions for other agents, they exaggerate tool capabilities. Researchers built a marketplace with 10 real Amazon or Instagram data tools, with 5 production models as sellers rewriting descriptions, and Claude Code and Codex as buyers selecting based only on descriptions; all 5 models exaggerated claims, including fields that tools returned on fewer than 50% of targets. When Gemini 3.8 Flash was the seller, the Amazon buyer success rate dropped from 89% to about 30%, close to the 28% of random selection. The authors suggest testing a model's seller behavior before it writes descriptions, requiring it to claim only real fields, and having the marketplace platform verify claims against tool outputs.

    #Papers/Research#Agents#English text is being prepared.

  9. Rohan PaulResearch
    AI score55

    Meta's new paper proposes RSD: agents can learn from unlabeled historical trajectories

    Meta's new paper proposes the RSD method, which lets AI agents review each step after knowing the task outcome and learn from historical attempts that no one labeled as right or wrong. On ScienceWorld, Qwen3-14B's success rate rose from 12.9% to 20.6%; on ALFWorld, Qwen3-8B reached 76.6%, surpassing all tested methods, including those that received rewards or standard answers. The authors recommend keeping agents' unlabeled run records (including failures) and regenerating the review notes each round, because when notes were frozen at round 3 the ALFWorld success rate stayed at 59.7%, while updating notes each round gave 67.9%.

    #Papers/Research#Agents#English text is being prepared.

  10. Rohan PaulResearch
    AI score54

    Harvard study: Paper reviews written by Claude Opus 4.8 outperform human reviews on all metrics

    A Harvard study used Claude Opus 4.8 to generate reviews for 1,124 papers and conducted a blind comparison against 134 expert reviews from rapidPeer. The results showed that the AI reviews outperformed the best expert reviews on all metrics tested by the authors. GPT-5.5 chose the AI review in all 132 decided pairs, and 3 former journal editors chose the AI review 87% of the time; erroneous statements about the papers accounted for 5.5% of AI reviews versus 15.9% of human reviews. Based on this, the authors suggest that teams relying on expert review can have AI complete a full first-round review, with humans making the final decision.

    Same story · 4 reports

    #Papers/Research#Evaluations / benchmarks#Reasoning

  11. Ethan MollickResearch
    AI score61

    Study shows Opus 4.8 comprehensively outperforms human experts in peer review

    A study covering 124 Harvard social science papers shows that Opus 4.8 outperforms human experts on every peer review metric, with papers ranging from statistical analysis and formal models to qualitative research. Ethan Mollick cited the study and noted that academia needs to think quickly about how to adapt scientific research practice to AI.

    #English text is being prepared.#Evaluations / benchmarks#Reasoning

  12. TMTPost · AGIResearch
    AI score63

    Anthropic report: Claude generates 180 million lines of Rocq proofs, advancing formal verification to the assembly level

    The Anthropic and Theorem Labs team released a report stating that Claude generated approximately 180 million lines of Rocq proof code for 239 program entry points across 123 open-source software libraries, and found and fixed hundreds of bugs. The research is based on the RISC-V instruction set and advances verification to the assembly instruction level; Anthropic says this is the first large-scale use of AI to formally verify production software at the assembly code level. The article also notes that formal verification can only prove that a program satisfies preset properties under specific assumptions and scope, and that the costs of AI use for attackers and defenders are not symmetric.

    Same story · 3 reports

    #Papers/Research#Safety / alignment#Agents

  13. The PaperResearch
    AI score53

    Academician Jia Weiping's team publishes research results on VR exercise intervention for obese adolescents and AI personalized exercise prescriptions

    Academician of the Chinese Academy of Engineering Jia Weiping reported at the 21st "Blue Pujiang" Academic Annual Conference that her team, together with Shanghai University of Sport, the Institute of Artificial Intelligence of Shanghai Jiao Tong University and others, conducted a clinical trial of VR exercise intervention for obese adolescents. After 8 weeks of intervention, the VR exercise group's body weight decreased by an average of 4.28 kilograms, close to the 5 kilograms in the real exercise group; liver enzyme indicators improved, blood lipids decreased, and cognitive speed increased by 21%. The team also developed a personalized exercise prescription system based on cardiorespiratory endurance, muscle strength and multi-omics assessment. In 16 weeks of combined aerobic and resistance training for patients with type 2 diabetes, participants showed reduced visceral fat, increased lower-limb skeletal muscle mass and endurance, and lower glycated hemoglobin; through AI multi-omics analysis, 27 serum differential proteins associated with body composition improvement were screened out, of which 16 were associated with changes in activity of cognitive brain regions.

    #Industry news#Multimodal#English text is being prepared.

  14. WallstreetcnResearch
    AI score64

    Goldman Sachs quantifies AI's speed of impact on employment: affected occupations see average job losses of about 25% over five years

    Goldman Sachs Global Investment Research published a report on October 22, using historical technology shock experience to estimate AI's impact on the labor market: if AI ultimately brings a 15% productivity gain, during the transition period up to about 9% of employees may temporarily leave their current jobs; in occupations where technology makes labor roles redundant, employment falls by an average of about 25% over five years, with photo lab technicians seeing job losses of nearly 60% over five years. The report also finds that each 1% of technology-driven productivity gain temporarily raises the peak unemployment rate by 0.3 to 0.4 percentage points, and that job contraction comes more from a rising separation rate than from slower hiring, so AI-related layoff announcements are better suited as a high-frequency indicator to watch.

    #Industry news#English text is being prepared.#Papers/Research

  15. cnBetaResearch
    AI score62

    The Lancet Digital Health publishes the FAHR-Face study: using facial photographs to estimate biological age and survival risk in cancer patients

    The Lancet Digital Health published a study in which the team developed the FAHR-Face model, trained on more than 40 million facial images, to estimate biological age and survival risk in cancer patients from facial features. After evaluation on data from about 43,000 cancer patients, 52% of the 122 patients classified by the model as high risk died within 90 days, while only 2% of the 975 patients classified as lowest risk died in the same period. The study notes that the method has a low barrier to entry, but causality has not been established, and how much additional predictive value it offers compared with existing clinical methods remains undetermined.

    #Papers/Research#Multimodal#English text is being prepared.

  16. Ayi AI NotesResearch
    AI score55

    University of the Basque Country and Cardiff University study: mainstream large models favor Japanese culture on 31,000 open-ended questions

    Researchers from the University of the Basque Country and Cardiff University tested mainstream large models with 31,000 open-ended cultural questions in 24 languages, finding that Gemini, ChatGPT, Claude, and Llama still disproportionately default to Japanese examples and paradigms after geographic restrictions are removed. The study locates the emergence point of the bias at the supervised fine-tuning (SFT) stage, arguing that the worldview at the pretraining stage is relatively balanced, while alignment training that pursues restraint and low conflict anchors Japanese-style cultural symbols into the models. The study also points out that the more aggressively a model is quantized and compressed, the more extreme this stereotypical preference becomes.

    #Papers/Research#Safety / alignment#English text is being prepared.

  17. AnthropicResearch
    AI score65

    Anthropic uses Claude to formally verify hundreds of open-source libraries

    Anthropic's Frontier Red Team blog introduces that, in collaboration with Theorem Labs, it used Claude to prove security properties for over a hundred real open-source libraries within weeks. Formal verification can prove that code is free of certain defects, but in the past it was only applicable to small codebases or required a large amount of expert time; Anthropic believes AI models are changing this situation. See the original link for the full research.

    Same story · 3 reports

    #Papers/Research#Safety / alignment#Agents

  18. 36Kr · AIResearch
    AI score77

    Google DeepMind beats Minecraft with four Gemini 4 Argon agents

    Google DeepMind published the paper "A team of Gemini 4 Argon agents beats Minecraft": four Gemini 4 Argon coding agents autonomously divided labor in a randomly generated Minecraft survival world, accumulating 605 deaths and writing 2,783 programs, defeating the Ender Dragon in 27 hours and 14 minutes. The experiment used Minecraft Java Edition 1.20.4, Easy difficulty, a random map, and no cheats; the agents read the environment via Mineflayer and wrote JavaScript to command characters, with roughly 24,000 model calls and 1.35 billion input tokens over the whole run, 91% of which came from context caching. The authors note that about 22% of game programs contained errors and the server restarted 6 times; the paper did not demonstrate stable completion through repeated tests on multiple random maps, and at discounted API prices the completion cost is estimated at about $520.

    Same story · 3 reports

    #Papers/Research#Agents#DeepMind

  19. Unite.AIResearch
    AI score77

    Anthropic says Claude completed formal proofs of security properties for 123 open-source libraries

    Anthropic released research on October 22, 2026, saying that in collaboration with Theorem, Claude generated machine-checkable formal proofs of security properties at the RISC-V assembly code level for 239 fuzz test entry points across 123 open-source libraries, and patched the vulnerabilities found. The proofs cover arbitrary inputs below 64 MB and modeled Linux system responses and memory layout, totaling about 180 million lines of Rocq, with a typical proof about 600,000 lines, corresponding to a median binary length of 18,000 RISC-V instructions. Anthropic said some proofs and the RISC-V and operating system semantics have been published on GitHub, but proofs that could expose unpatched vulnerabilities will be withheld; Theorem is offering the same process free of charge to eligible open-source projects through the OSS Verify program.

    Same story · 3 reports

    #Papers/Research#Safety / alignment#Agents

  20. TechCrunch AIResearch
    AI score64

    MIT study: 13 Claude Opus 4.7 agents perform better after playing autonomously on a virtual deserted island for 30 hours

    MIT researchers had 13 Eko agents based on Claude Opus 4.7 explore a simulated deserted island without any assigned tasks for 30 hours; afterward, in designated challenges, these agents generally outperformed comparable agents that had never been on the island. The paper has been published on arXiv and has not yet undergone peer review. The study also found that 6 agents figured out on their own that letting go mid-jump allows them to throw blocks farther, an ability the designers had not anticipated; UC Berkeley cognitive scientist Eunice Yiu considers the results impressive but hopes to see more evidence about the cognition behind the behavior.

    #Papers/Research#Agents#arXiv