Source URL: https://www.forbes.com/sites/daveywinder/2020/02/28/google-confirms-new-ai-tool-scans-300-billion-gmail-attachments-every-week/
Source: Hacker News
Title: I can’t stand Google scanning my emails anymore
Feedly Summary: Comments
AI Summary and Description: Yes
Summary: Google is leveraging advanced deep learning AI models to enhance the security of Gmail by effectively detecting and blocking malicious attachments. The application of these models reflects a significant advancement in email security technology, as Google claims it blocks 99.9% of malicious Gmail attachments.
Detailed Description:
– Google has undertaken a significant initiative to improve the security of Gmail, particularly focusing on the detection of malware hidden in email attachments.
– Malware in documents, such as macro viruses, poses a substantial threat, constituting 58% of all malware targeting Gmail users.
– With deep learning AI, Google aims to tackle the ongoing issue of attachment malware.
Key Points:
– **Deep Learning Implementation**: The deep learning AI models have been deployed to enhance the capabilities of existing machine learning models, yielding an increase in detection coverage of malicious Office documents by 10%.
– **High Volume Processing**: Google’s systems process around 300 billion Gmail attachments weekly, with 63% of the blocked documents being unique daily.
– **Effectiveness**: The new AI models have improved detection rates for specific types of attacks, particularly those driven by botnets, boasting a 150% increase in identifying malicious documents during bursty attacks.
– **Technological Foundation**: The foundation of this deep learning approach includes the TensorFlow platform, with custom document analyzers for different file types that help in identifying attack patterns and de-obfuscating document content.
– **Industry Relevance**: The advancements in Gmail security with deep learning not only address ongoing challenges but also set a precedent in the email security landscape, providing an essential reference for organizations concerned about protecting sensitive information from evolving malware threats.
These developments have meaningful implications for security and compliance professionals as they represent a proactive approach to safeguarding email communications, an area often exploited for malware distribution. The continuous evolution of malware necessitates innovative defenses, highlighting the importance of adopting advanced AI technologies in security infrastructures.