search
Microsoft Research
Trends
- 1OpenAI and Microsoft researchers warn of AI 'doom loop' consuming the web▼‘Doom Loop’: OpenAI and Microsoft Admits LLMs Are Destroying the Web and Built on Theft
Researchers at OpenAI and Microsoft have published a paper describing a 'doom loop' in which large language models, trained on data scraped largely without consent from human creators, degrade the open web and eventually poison their own training data. The paper reportedly acknowledges that people will come to see the models' wholesale ingestion of creative work as an unprecedented act of theft, reigniting debate over copyright and the sustainability of generative AI.
- 2NASA turns to AI to mine decades of scientific data▼At NASA, AI is helping scientists unlock discoveries hidden in decades of data
NASA scientists are using artificial intelligence to comb through decades of archived data, surfacing discoveries that earlier analysis missed. A report by Microsoft highlights how machine learning tools are being applied to the space agency's vast historical datasets, helping researchers find new patterns and insights in old missions. The work illustrates a growing trend of AI-assisted science at major research institutions.
- 3
New reporting highlights phishing as a growing threat, with exposure rates approaching 70% across key US industries and security teams being urged to strengthen their defenses. Separately, researchers warn of AgtaBackup RAT, which uses fake Microsoft Store pages and remote monitoring tools to hijack Windows systems.
- 4Microsoft unveils Quine, an AI research system for biology●Introducing Quine: An AI research system designed for the complexity of biology
Microsoft has introduced Quine, a new AI research system built to handle the complexity of biology. The company positions the system as a tool for scientific research in the life sciences, where data and systems are unusually intricate. Details about its capabilities and availability have not yet been widely reported.
- 5Researcher Says Flaw Exposed 17 Trillion Microsoft Records●I could've accessed 17T Microsoft records
A security researcher has published a write-up describing a vulnerability that could have exposed roughly 17 trillion Microsoft records. The blog post, which has drawn attention on Hacker News, walks through how the flaw was found and reportedly reported to Microsoft. It is fueling discussion about cloud data security and the scale of exposure such bugs can reach.
- 6Microsoft Research Unveils Quine, a Multimodal Biology Model▼Microsoft Research Debuts Quine, a Multimodal World Model of Biology
Microsoft Research has introduced Quine, a multimodal world model of biology. According to a report by Unite.AI, the system is designed to integrate different types of biological data into a unified model. Details about its capabilities, benchmarks and intended applications remain limited in the initial coverage, and independent expert reaction has not yet been reported.
- 7New Windows NCSI proxy authentication flaw detailed by researchers▼Microsoft Windows NCSI Cross-Context Proxy Authentication Coercion - ZDI-26-708 - Part 1: https:// pgj11.com/posts/Windo
Security researchers have published a two-part technical write-up of a Windows vulnerability tracked as ZDI-26-708, described as a cross-context proxy authentication coercion in the Network Connectivity Status Indicator (NCSI). The disclosure is circulating among security professionals sharing the detailed analysis. Details on affected versions and patches remain unclear from the discussion so far.
- 8Microsoft details Zimbra flaw allowing code execution via email●Zimbra CVE-2026-73570 lets one crafted email run code. Microsoft details the Zimbra command injection attacks: web shell
Microsoft researchers have detailed attacks exploiting a Zimbra command injection vulnerability, CVE-2026-73570, in which a single crafted email is enough to run code on the mail server. The documented attacks involve deploying web shells, gaining root access, and stealing cryptographic keys. Security teams running Zimbra are being urged to patch and review their servers for signs of compromise.
- 9Microsoft rates Security Copilot prompt injection risk as Low●MSRC assessed an indirect prompt injection into Security Copilot as Low severity because consequential action still requ
Microsoft's Security Response Center assessed an indirect prompt injection into Security Copilot as Low severity, reasoning that consequential action still required downstream automation or human approval. Security researchers are pushing back, arguing that rationale becomes untenable as AI systems are increasingly designed to perceive, reason, and act autonomously, with less human oversight built in.
- 10Microsoft Launches Quine AI for Scientific Discovery▼Microsoft Pushes AI Deeper Into Scientific Discovery With Quine
Microsoft has introduced Quine, an AI system aimed at advancing scientific discovery. According to HPCwire, the company is pushing artificial intelligence deeper into research workflows, positioning Quine as a tool to accelerate experimentation and analysis in the sciences. Details about its capabilities and target research fields remain limited, and reaction from the scientific community is still emerging.
- 11Microsoft details NeedyMantis malware used in targeted attacks●Posted yesterday, if you missed this. Microsoft: NeedyMantis: Unpacking a post-compromise malware family used in targete
Microsoft has published an analysis of NeedyMantis, a malware family deployed after attackers have already breached a network, with use in targeted operations against specific victims. The report breaks down how the malware behaves once inside a compromised environment. Security researchers and practitioners are sharing the findings, warning organisations to review the indicators of compromise Microsoft disclosed.
- 12Microsoft-linked network spotted announcing IP space from Oslo●ASN: AS8075 Location: Oslo, NO Added: 2026-09-25T18:01 # shodansafari # infosec
Autonomous System 8075, the network operated by Microsoft, was observed announcing IP address space located in Oslo, Norway. The observation was logged and shared on 25 September 2026 with the cybersecurity community under the tag #shodansafari. This type of sighting is used by security researchers to track how large cloud and technology providers extend their network presence into new regions and data centre locations.