search
Microsoft Research
Trends
- 1
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.
- 2Researcher 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.
- 3
Peter Lee, who has led Microsoft's research and science operations, is stepping down from his role as president. The news was first reported by The Information. Lee has been a central figure in Microsoft's research organization, overseeing work spanning artificial intelligence and other scientific fields. Details on his departure and a successor have not yet been announced.
- 4Microsoft 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.
- 5Microsoft 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.