Microsoft's New AI Security Tool Can Spot Malware Early - and Even Reverse Engineer It to Crack the Code
MICROSOFT, AUG 5 – Project Ire autonomously reverse engineers malware with 90% accuracy and a 2% false positive rate, aiming to assist human researchers and enhance threat detection in Microsoft Defender.
- On Tuesday, Microsoft announced Project Ire, prototype AI for autonomous malware detection via reverse engineering of software files.
- Amid evolving attack techniques, Project Ire uses autonomous reverse engineering and large language models to help the IT security sector leverage AI against concealed threats.
- In tests involving nearly 4,000 files, Project Ire correctly flagged roughly a quarter of malware and identified 90% of files in a Windows driver dataset, with only 2% false positives.
- Microsoft will integrate Project Ire into Microsoft Defender as a binary analyzer, which already produced an automatic blocking decision against malware tied to an elite hacking group.
- Looking ahead, the team plans to scale Ire’s speed and accuracy for first-encounter classification, emphasizing its role in assisting overburdened human analysts amid rising AI adoption in cybersecurity.
Insights by Ground AI
Does this summary seem wrong?
11 Articles
11 Articles
Microsoft presented today the Ire Project, a prototype of Artificial Intelligence (IA), capable of identifying and classifying malware ('malware') independently, helping to strengthen existing security measures. The project results from the joint work of several Microsoft departments, integrating advanced language models, reverse engineering tools and binary analysis. Due to the application of IA to identify and classify 'malware' from there
·Funchal, Portugal
Read Full ArticleWith Project Ire, Microsoft introduces an AI system that automatically analyzes software files and evaluates whether they contain malware.The article Microsoft introduces new AI for malware detection with Project Ire first appeared on THE-DECODER.de.
·Germany
Read Full ArticleCoverage Details
Total News Sources11
Leaning Left2Leaning Right1Center1Last UpdatedBias Distribution50% Left
Bias Distribution
- 50% of the sources lean Left
50% Left
L 50%
C 25%
R 25%
Factuality
To view factuality data please Upgrade to Premium